United States Of America v. Bazaarvoice, Inc.

Filing 244

MEMORANDUM OPINION (PUBLIC REDACTED VERSION). The government has prevailed in the liability phase of this case. The parties shall file a Joint Statement Regarding Remedy Phase on January 17, 2014, and attend a Case Management Conference on January 22, 2014, at 9 a.m. in Courtroom 2, 450 Golden Gate Avenue, San Francisco, California, to discuss the remedy phase of this case. Signed by Judge William H. Orrick on 01/08/2014. (jmdS, COURT STAFF) (Filed on 1/8/2014)

Download PDF
1 2 3 4 UNITED STATES DISTRICT COURT 5 NORTHERN DISTRICT OF CALIFORNIA 6 7 UNITED STATES OF AMERICA, Case No. 13-cv-00133-WHO Plaintiff, 8 v. MEMORANDUM OPINION 9 10 BAZAARVOICE, INC., Defendant. United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 PUBLIC REDACTED VERSION CONTENTS 1 2 3 4 INTRODUCTION ............................................................................................................................. 4 SUMMARY OF FINDINGS ............................................................................................................ 5 FINDINGS OF FACTS ................................................................................................................... 10 5 I. THE MERGING PARTIES AND JURISDICTION .............................................. 10 6 II. OVERVIEW OF R&R AND THE INDUSTRY .................................................... 11 7 A. R&R Platforms ............................................................................................ 11 8 B. Product Comparison Between Bazaarvoice and PowerReviews ................ 14 9 C. Selling R&R Platforms................................................................................ 16 10 D. The Broader Social Commerce and eCommerce Economy ........................ 19 United States District Court Northern District of California 11 III. 12 WHY BAZAARVOICE PURCHASED POWERREVIEWS ................................ 21 A. PowerReviews Was Bazaarvoice’s Only Real Commercial Competitor ... 21 B. Bazaarvoice’s Pre-Acquisition Rationales For The Merger Were Anticompetitive ........................................................................................... 29 C. Bazaarvoice’s Trial Rationale for the Acquisition ...................................... 36 13 14 15 16 IV. 17 THE RELEVANT MARKET ................................................................................. 41 A. B. 19 The United States Is the Relevant Geographic Market ............................... 51 C. 18 The Relevant Product Market is R&R Platforms ........................................ 41 Economic Analysis Confirms that R&R Platforms Sold to United States Customers Is the Relevant Market .............................................................. 53 20 21 22 23 24 V. ANALYSIS OF GOVERNMENT’S PRIMA FACIE CASE AND BAZAARVOICE’S RESPONSE ............................................................................ 62 A. Bazaarvoice’s market share establishes a prima facie violation of Section 7 ..................................................................................................................... 62 B. Bazaarvoice cannot rebut the government’s prima facie case: entry or repositioning from other firms has not, and will not, counteract Bazaarvoice’s increased market power ....................................................... 73 C. It is probable that the transaction will substantially lessen competition and result in higher prices ................................................................................ 102 25 26 27 28 2 D. 1 2 3 Bazaarvoice has not demonstrated that merger-specific efficiencies will overcome the anticompetitive effects of the merger ................................. 118 DISCUSSION ............................................................................................................................... 122 4 THE HIGH CONCENTRATION OF THE RELEVANT MARKET IS KEY TO THIS CASE ........................................................................................................... 123 5 A. Relevant product market ........................................................................... 124 6 B. Relevant geographic market ...................................................................... 126 7 I. II. 8 BAZAARVOICE CANNOT REBUT THE GOVERNMENT’S PRIMA FACIE CASE ..................................................................................................................... 127 A. Bazaarvoice’s percentage share of the relevant market exceeds 50 percent and establishes a prima facie Section 7 violation ...................................... 127 B. 9 Bazaarvoice failed to rebut the government’s prima facie case ................ 131 10 11 United States District Court Northern District of California III. APPLICATION OF ANTITRUST LAW TO DYNAMIC MARKETS ............... 140 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 3 INTRODUCTION 1 1 Bazaarvoice, the unquestioned market leading provider of Ratings and Reviews platforms 2 3 (“R&R”) to companies involved in online commerce (referred to in this Opinion as “eCommerce”) 4 in the United States, acquired PowerReviews, its primary competitor, on June 12, 2012. The 5 United States Department of Justice launched an antitrust investigation two days later and filed 6 suit on January 10, 2013, alleging that the transaction violated Section 7 of the Clayton Act, 15 7 U.S.C. § 18. This Memorandum Opinion explains why the government has prevailed in the 8 liability phase of this case. Section 7 of the Clayton Act prohibits mergers whose effect “may be substantially to 9 lessen competition, or to tend to create a monopoly.” 2 15 U.S.C. § 18. The Supreme Court has 11 United States District Court Northern District of California 10 explained that “Congress used the words ‘may be substantially to lessen competition’ [ ], to 12 indicate that its concern was with probabilities, not certainties.” Brown Shoe Co. v. United States, 13 370 U.S. 294, 323 (1962). Consequently, to establish a violation of Section 7, the government 14 need not prove that the merger has resulted in higher prices or had other anticompetitive effects. 15 Rather, the government must show a “reasonable likelihood” of anticompetitive effect in the 16 relevant market. United States v. Penn-Olin Chem. Co., 378 U.S. 158, 170 (1964) (noting that a 17 Section 7 violation is shown when “the ‘reasonable likelihood’ of a substantial lessening of 18 competition in the relevant market is shown”) (citing United States v. E. I. du Pont de Nemours & 19 1 20 21 22 The Court provided this Memorandum Opinion to the parties in final form on January 8, 2014. Public release was delayed to ensure that no confidential information was disclosed. The Court separately issued an order granting various third parties’ motions to seal their documents used as trial exhibits. Based on the motions to seal and input from the parties, confidential information has been redacted from this opinion, reflected by the insertion of black boxes covering the redacted text. 23 2 24 25 26 27 28 Section 7 of the Clayton Act states, in relevant part: No person engaged in commerce or in any activity affecting commerce shall acquire, directly or indirectly, the whole or any part of the stock or other share capital and no person subject to the jurisdiction of the Federal Trade Commission shall acquire the whole or any part of the assets of another person engaged also in commerce or in any activity affecting commerce, where in any line of commerce or in any activity affecting commerce in any section of the country, the effect of such acquisition may be substantially to lessen competition, or to tend to create a monopoly. 15 U.S.C. § 18. 4 1 Co., 353 U.S. 586, 592 (1957)); see also F.T.C. v. Warner Commc’ns Inc., 742 F.2d 1156, 1160 2 (9th Cir. 1984) (“It is well established that a section 7 violation is proven upon a showing of 3 reasonable probability of anticompetitive effect.”); United States v. Oracle Corp., 331 F. Supp. 2d 4 1098, 1109-10 (N.D. Cal. 2004) (“Section 7 does not require proof that a merger or other 5 acquisition [will] cause higher prices in the affected market. All that is necessary is that the 6 merger create an appreciable danger of such consequences in the future.”). Conversely, while a 7 “reasonable likelihood” of anticompetitive effect establishes a Section 7 violation, an “ephemeral 8 possibility” of anticompetitive effect is insufficient. United States v. Marine Bancorporation, Inc., 9 418 U.S. 602, 622-23 (1974) (“it is to be remembered that [Section] 7 deals in ‘probabilities,’ not 10 United States District Court Northern District of California 11 ‘ephemeral possibilities’”) (citing Brown Shoe, 370 U.S. at 323). The government has the burden of proving a Section 7 violation by a preponderance of the 12 evidence. Concord Boat Corp. v. Brunswick Corp., 207 F.3d 1039, 1058 (8th Cir. 2000); Oracle 13 Corp., 331 F. Supp. 2d at 1109; United States v. Sungard Data Sys., Inc., 172 F. Supp. 2d 172, 180 14 (D.D.C. 2001). It is authorized to seek an injunction to enforce Section 7. 15 U.S.C. § 25. 15 The trial of this matter ran from September 23, 2013, through October 15, 2013. After 16 listening to the testimony of 40 witnesses, reading more than 100 depositions, reviewing 980 17 exhibits, and considering the memoranda and arguments of counsel, the Court finds that 18 Bazaarvoice is liable for violating Section 7 of the Clayton Act. 19 This necessarily lengthy Opinion will first summarize the findings that lead to the 20 conclusion that Bazaarvoice has violated Section 7. Following that is a more detailed discussion 21 of the facts found by the Court, starting with industry, company, and product backgrounds, then 22 moving to Bazaarvoice’s acquisition of PowerReviews, the relevant market, Bazaarvoice’s market 23 share, and the facts that show why it is probable that the merger will have an anticompetitive 24 effect and why Bazaarvoice did not overcome the government’s prima facie case. The Opinion 25 concludes with a discussion of the legal issues raised. 26 SUMMARY OF FINDINGS 27 R&R platforms combine software and services to enable manufacturers and retailers, 28 among other companies, to collect, organize, and display consumer-generated product reviews and 5 1 ratings online. These reviews and ratings are displayed on the page of a commercial website on 2 which a consumer can purchase the product discussed. The purpose of R&R is to drive up the 3 “conversion rate” of consumers—the rate at which visitors to a website purchase products from 4 the website. R&R is a valuable asset for retailers and manufacturers to feature on their websites 5 because it tends to increase sales, decrease product returns, improve a website’s ranking by 6 Internet search engines, and provide valuable information regarding consumer sentiment. While 7 R&R platforms provide just one type of user-generated-content (“UGC”) that retailers and 8 manufacturers can use on their eCommerce sites, it is uniquely effective in increasing the 9 conversion rate. The Court finds that R&R is in the dynamic and fast-evolving eCommerce industry, but 10 United States District Court Northern District of California 11 the consistent testimony of customers and other evidence establish that R&R is critical or, as 12 PowerReviews’s CEO called it, “table stakes” for companies selling product over the internet. 3 It 13 is a separate product market for antitrust purposes. The geographic market is the United States 14 because, among other reasons, the realities of marketing and servicing R&R platforms locally, 15 cultural differences in moderating the user-generated content, and the difficulties of incorporating 16 different SKUs distinguish the United States from the international market. The evidence that Bazaarvoice and PowerReviews expected the transaction to have 17 18 anticompetitive effects is overwhelming. Bazaarvoice recognized PowerReviews as its only real 19 commercial competitor, and vice versa. Exhibit after exhibit manifest that Bazaarvoice and 20 PowerReviews viewed themselves as operating in a “duopoly” and that removing 21 PowerReviews—an established company offering R&R at cheaper price points with a significant 22 customer base as measured by the Internet Retailer 500 (“IR500”), a well-recognized list of the 23 leading 500 internet retailers—would eliminate Bazaarvoice’s only meaningful commercial 24 competitor. The R&R platforms market was concentrated before the merger, and the acquisition by 25 26 Bazaarvoice of its primary commercial competitor establishes the government’s prima facie case. 27 3 28 The Court takes “table stakes,” in this context, to mean an offering companies must provide to effectively sell product over the internet. 6 1 The government’s expert, Dr. Carl Shapiro, testified convincingly that Bazaarvoice’s acquisition 2 of PowerReviews is likely to have anticompetitive effects. 3 Much of what Bazaarvoice refers to now as its “competitive strengths” it used to call, 4 accurately, significant barriers to entry. When the network effect of syndication and switching 5 costs are combined with reputation and Bazaarvoice’s market share, it has a substantial advantage 6 over its smaller competitors and any company that is thinking about entering the market. 7 Bazaarvoice was unable to rebut the government’s prima facie case. Besides disputing the 8 market and market shares proposed by the government and ultimately found by the Court, it 9 argues that the real purpose of its purchase of PowerReviews was to allow it to compete more effectively in advertising and “big data,” which is in a much broader market than R&R. That may 11 United States District Court Northern District of California 10 be Bazaarvoice’s ultimate goal, but it has no intention of dropping R&R, around which the 12 company is built. The purchase of PowerReviews provides “breathing space” for Bazaarvoice in 13 R&R while it prepares to compete in the broader market. While the commercial strategy may 14 make sense, Bazaarvoice is not allowed to accomplish it by violating the Clayton Act, as it did by 15 acquiring PowerReviews. 16 Bazaarvoice argues that PowerReviews was a weak competitor and points to the growing 17 number of companies in the social commerce space, several of which offer R&R solutions, and in- 18 house solutions as substitutes for PowerReviews’s place in the R&R market. In-house solutions 19 can be affordable and acceptable alternatives for some. But due to cost, on-going maintenance 20 and development requirements, and company priorities, the number of companies going in-house 21 has not appreciably increased over time. And while smaller companies with basic needs can find a 22 variety of lower cost solutions than Bazaarvoice for R&R software, these other providers and in- 23 house solutions have not made much of a dent in Bazaarvoice’s market share. It took 24 PowerReviews seven years to develop its competitive place in the R&R market, with around 80 25 customers in the IR500 at the time it was purchased. Before the acquisition, Bazaarvoice 26 estimated that it would cost between $32 million and $50 million to compete successfully with 27 PowerReviews for its IR500 clients. It is not probable that current competitors or potential market 28 entrants can threaten the merged entity to the same degree that PowerReviews did, as shown by 7 1 the competitors’ lack of success. It is unlikely that PowerReviews will be replaced by the existing 2 R&R competitors in the next two years, the time frame in which the Court evaluates the likely 3 effects of the merger. 4 Bazaarvoice points to behemoths like Amazon and Google as potential rapid entrants in 5 R&R. Amazon already can be considered a competitor in the market—its internal R&R solution 6 accounts for 28 percent of the R&R market by one measurement used by Dr. Shapiro. Many of 7 Bazaarvoice’s customers also use Amazon’s websites to sell their products, but because they are 8 also natural competitors with Amazon, having a separate R&R platform makes sense for most 9 companies. On the other hand, Google is seen as more complementary to Bazaarvoice than competitive. No credible evidence was presented to show that any rapid entrants will join this 11 United States District Court Northern District of California 10 product market in the next two years to meaningfully restrain Bazaarvoice’s anticompetitive 12 market power. 13 Finally, Bazaarvoice pointed out that none of the 104 customers whose depositions are part 14 of the record complained that the merger has hurt them. However, it would be a mistake to rely on 15 customer testimony about effects of the merger for several reasons. Bazaarvoice’s business 16 conduct after the merger was likely tempered by the government’s immediate investigation; the 17 customers were not privy to most of the evidence presented to the Court, including that of the 18 economic experts; many of the customers had paid little or no attention to the merger; and each 19 had an idiosyncratic understanding of R&R based on the priorities of their company and their 20 different levels of knowledge, sophistication, and experience. Moreover, the basis of 21 Bazaarvoice’s (and other R&R platforms vendors’) pricing proposals in negotiations are opaque to 22 customers. Price is negotiated individually with each customer, based on the customer’s need and 23 Bazaarvoice’s ability to customize its product and sell additional features to the customer. It is 24 difficult for those customers to discern what is actually happening in the market. 25 The analysis in this case is complicated by several factors worth mentioning at the outset. 26 First, there are no generally accepted market share statistics covering the sales of R&R solutions 27 or social commerce solutions and no perfect way to measure market shares. Each party presented 28 analyses that had unavoidable flaws. Even so, the Court finds that Dr. Shapiro’s approaches were 8 1 more persuasive than those of Bazaarvoice’s economist, Dr. Ramsey Shehadeh. Second, testimony from customers showed that there is support from some customer for 2 3 virtually every claim made in this case about the value (or lack thereof) of different R&R platform 4 features because the needs and priorities of each company are different. 5 Third, the potential for witness bias was greater in this case than most. Many of the party 6 witnesses, and some of the customers, had significant personal financial interests in the outcome 7 of this case. 4 Third-party customers had to testify about their market strategy in front of a vendor 8 with whom most would be negotiating within a short time. Fourth, as noted earlier, Bazaarvoice’s employees and officers undoubtedly considered 9 how its post-merger conduct would appear to the government and the Court, especially given the 11 United States District Court Northern District of California 10 government’s immediate scrutiny of the transaction. 12 While Bazaarvoice fought against every material argument of the government, its defenses 13 were often undermined by pre-acquisition statements from its and PowerReviews’s executives. Its 14 portrayal of PowerReviews as a weak and unworthy competitor was belied by the plethora of 15 documents showing that, prior to the merger, Bazaarvoice considered PowerReviews its strongest 16 and only credible competitor, that the two companies operated in a duopoly, and that 17 Bazaarvoice’s management believed that the purchase of PowerReviews would eliminate its only 18 real competitor. Similarly, Bazaarvoice’s attempt to emphasize the strength of existing 19 competitors at trial was belied by the testimony of the competitors and Bazaarvoice officials prior 20 to trial as well as the lack of success of those competitors in attracting R&R customers. 21 As the Supreme Court has instructed, the Court is dealing with probabilities, not 22 23 24 25 26 27 28 4 Witnesses Stephen Collins (Bazaarvoice’s CEO and member of the Board of Directors), Brett Hurt (Bazaarvoice’s co-founder and former CEO and current Vice-Chairman of the Board of Directors), Thomas Meredith (Chairman of Bazaarvoice’s Board), Brant Barton (Bazaarvoice’s co-founder), Michael Osborne (Bazaarvoice’s former Chief Revenue Officer), and Kenneth Comée (PowerReviews’s former CEO) all own a substantial amount of Bazaarvoice stock and are financially interested in the outcome of this case. Collins, Hurt, and Meredith have also been sued in their personal capacity in a private lawsuit in Texas state court alleging that they knew that the acquisition would lessen competition and create a monopoly, and failed to disclose the anticompetitive nature of the acquisition. Third-party witness Christian Friedland of Build.com is an investor in Austin Ventures, one of Bazaarvoice’s largest shareholders. 9 1 certainties, to determine if the merger is likely to create anticompetitive effects. Given the 2 overwhelming market share Bazaarvoice acquired when it purchased PowerReviews, the stark pre- 3 merger evidence of anticompetitive intent and the merger’s likely effects, coupled with the actual 4 lack of impact competitors have made since the merger, the government established the Section 7 5 violation. The Court finds that it is probable that Bazaarvoice’s acquisition of PowerReviews will 6 have an anticompetitive effect over the next two years absent intervention by the Court. The evidence at the trial focused on liability. In light of the findings of the Court, the 7 8 government established that, at law, it would be entitled to an injunction that requires Bazaarvoice 9 to divest PowerReviews. That is not a simple proposition 18 months after the merger, and the Court would benefit from hearing the parties’ perspectives on the appropriate remedy. A Status 11 United States District Court Northern District of California 10 Conference is set for January 22, 2014, at 9:00 a.m., in Courtroom 2 to discuss the procedure for 12 the remedy phase of this litigation. The parties should confer and file a Joint Statement Regarding 13 Remedy Phase four days prior to the Status Conference to propose a procedure by which the 14 remedy should be determined. 15 16 FINDINGS OF FACTS I. THE MERGING PARTIES AND JURISDICTION 17 1. Bazaarvoice, a publicly traded Delaware corporation headquartered in Austin, Texas, was 18 founded in 2005. The vast majority of Bazaarvoice’s customers purchase its R&R 19 platform, which accounts for the vast majority of Bazaarvoice’s revenue. Trial Tr. 5 20 881:11-18 (Collins); Trial Tr. 1424:13-16 (Meredith). During its 2012 fiscal year, 21 Bazaarvoice earned approximately $106.1 million in revenue. 22 2. PowerReviews was Bazaarvoice’s closest commercial competitor in the provision of R&R 23 platforms in the United States when it was acquired on June 12, 2012. It was a privately 24 25 26 27 28 5 The Court uses abbreviations “GX” and “DX” to refer to the government’s trial exhibits and Bazaarvoice’s trial exhibits, respectively. “Trial Tr.” refers to trial testimony. “Dep.” refers to pre-trial deposition testimony. “PFF” refers to the government’s post-trial plaintiff's proposed findings of fact. Dkt. No. 235. “DFF” refers to Bazaarvoice post-trial proposed findings of fact. Dkt. No. 239. 10 1 held Delaware corporation headquartered in San Francisco, California, founded the same 2 year as Bazaarvoice. GX318 (Def.’s Answer ¶ 12). The vast majority of PowerReviews’s 3 customers purchased its R&R platform, and R&R accounted for the vast majority of 4 PowerReviews’s revenue. Trial Tr. 420:17-23, 442:6-16 (Luedtke). During the 2011 5 calendar year, the company earned approximately $11.5 million in revenue. GX318 6 (Def.’s Answer ¶ 12). 3. Bazaarvoice sells its products throughout the United States. Its sales constitute a regular, 7 8 continuous, and substantial flow through interstate commerce, and have a substantial effect 9 upon interstate commerce. It transacts business and has offices within the Northern District of California. 10 4. The Court has personal jurisdiction over Bazaarvoice and subject matter jurisdiction under United States District Court Northern District of California 11 Section 15 of the Clayton Act, 15 U.S.C. § 25. 12 13 II. OVERVIEW OF R&R AND THE INDUSTRY 14 15 16 A. R&R Platforms 5. R&R platforms are one of many social commerce tools. “Social commerce” is “a very 17 broad term” that refers generally to the engagement or interaction of consumers as part of 18 their shopping or buying process, typically in a digital medium such as a website. Trial Tr. 19 1852:1-8 (Meisels). To determine the appropriate product market and understand the 20 effect of the merger, one must first understand what R&R platforms are and how they fit 21 into the social commerce tools industry. 22 6. The software component of a R&R platform provides a user interface and review form for 23 the collection and display of user-generated content on the product page of a commercial 24 website where the product can be purchased. Most review forms prompt consumers to rate 25 a product on a five-star scale and offer consumers an option to write an open-ended review 26 of the product. Other forms also allow consumers to rate products along several 27 dimensions (e.g., product appearance, ease of assembly, value). There are various features 28 offered by R&R software and social commerce suppliers such as syndication, moderation, 11 1 attribute ratings, and data analytics. R&R platforms vary in their level of sophistication, 2 features offered, and other available capabilities that may be bundled with the platform or 3 sold separately. 7. R&R platforms increase the likelihood a consumer will purchase a R&R customer’s 5 product on the customer’s website. Consumers find R&R useful because it provides 6 authentic information regarding another consumer’s experience with a particular product. 7 Feedback from other consumers can help a prospective buyer make a more informed 8 purchasing decision. Even negative reviews have a positive impact on sales because they 9 confirm the authenticity of all R&R. GX82 (Barton Dep. 41:21-43:12); GX928 at -926; 10 Trial Tr. 1195:7-16 (Friedland/Build.com) (“There is a 30 percent lift associated with -- 11 United States District Court Northern District of California 4 that we’ve measured . . . if you sell product that has zero reviews versus a product that has 12 a handful of reviews. Even if they’re negative, you sell more, oddly enough.”). 13 8. In addition to the basic ratings and reviews components of R&R platforms, sophisticated 14 R&R platforms also allow manufacturers to share, or “syndicate,” ratings and reviews with 15 their retail partners. Through the syndication network, retailers can display user-generated 16 content that was originally collected by a product’s manufacturer. For example, a retailer’s 17 website, such as Staples.com, can display ratings and reviews for a product, such as a 18 Canon printer, that were collected by Canon on Canon.com. Syndication is an important 19 selling feature for any R&R platform provider. GX82 (Barton Dep. 44:14-45:20); 20 (reviewed in camera); GX146 (Home Depot Dep. 62:4-9); see 21 also GX537 at -625. Syndication is of particular interest to manufacturers (also referred to 22 as “brands”) and is increasingly popular among retailers because it allows them to obtain 23 more content than they could independently. 24 9. Although each R&R customer has unique interests and priorities, as a general proposition, 25 manufacturers and retailers both benefit from the ability to display more reviews at the 26 point of sale. The increase in R&R content helps retailers sell more products and further 27 reduce return of products. Because R&R increases conversion rates, manufacturers also 28 see a benefit in the form of increased sales of their products through retailers. 12 1 10. Moderation is another key R&R feature. After receiving a review, but before the review is 2 posted on the manufacturer’s or retailer’s website. the R&R platform provider applies 3 software algorithms to scan the submission for inappropriate or fraudulent content. After 4 the automated scan, a human moderator (either provided by the R&R provider, a third 5 party, or in-house personnel) examines each submission to ensure that it complies with a 6 particular customer’s moderation standards. Moderation ensures that the R&R is credible, 7 not offensive, and otherwise complies with the company’s business rules. 8 11. Some R&R platforms also include analytics software that manufacturers and retailers use to analyze information collected from R&R. With these tools, manufacturers and retailers 10 can track and analyze real-time consumer sentiment. Manufacturers and retailers can use 11 United States District Court Northern District of California 9 this information to identify product design defects, make product design decisions, or 12 identify consumers for targeted marketing efforts. R&R provides valuable data about 13 consumer preferences and behavior, which retailers and manufacturers can use to make 14 inventory purchasing or product design decisions. 15 12. Internet search engines, such as Google, Yahoo, and Bing, generally rank websites with 16 content that is fresh and unique higher on the search engine’s results page than websites 17 with less fresh and unique content. Search engine optimization (“SEO”) is the process of 18 making a website more “visible” to a search engine, leading to a higher search engine 19 rankings. 20 13. R&R must be visible to search engines to assist with SEO. Because search engines prefer 21 fresh and unique content, when R&R is detected and indexed by a search engine, the 22 underlying product pages will likely receive a higher ranking on the search engine’s results 23 page, leading to increased traffic to those product pages. PowerReviews used to, and 24 Bazaarvoice does, provide SEO functionality that enables R&R to be detected and indexed 25 by a search engine. GX484 at -017; GX485 at -124-125. Syndication further helps 26 improve a website’s visibility because “authentic content is prioritized” by search engines. 27 Trial Tr. 826:13-827:1 (Collins); see also GX1248. R&R has a “very positive” effect on 28 13 1 SEO. GX141 (Golfsmith Dep. 36:1-11); Trial Tr. 1218:19-22 (Maki/Golfsmith); see also 2 GX195 (Wine.com Dep. 30:7-30:14); GX125 (Clorox Dep. 111:23-112:18). 3 14. Retailers and manufacturers can conduct tests to confirm the extent to which a R&R 4 platform increases the conversion rates on their website. GX168 (Onlineshoes.com Dep. 5 29:18-31:3). Online retailers confirm that R&R increases conversion on their websites. 6 GX193 (Walgreens Dep. 40:14-41:5); GX179 (REI Dep. 48:12-49:8); GX109 (Blinds.com 7 Dep. 38:17-39:3); GX195 (Wine.com 29:15-30:6). Consumers often seek out R&R online 8 even when making in-store purchases, so R&R can also have a positive effect on a 9 retailer’s in-store sales in addition to its online sales. GX662 at -267-268; GX141 (Golfsmith Dep. 36:22-23, 36:25-37:7, 37:9-11); Trial Tr. 1218:23-1219:1 11 United States District Court Northern District of California 10 (Maki/Golfsmith); GX131 (Dillard’s Dep. 70:2-4, 70:7-71:6); GX146 (Home Depot Dep. 12 52:25-53:4, 53:9-12); 13 15. Both Bazaarvoice and PowerReviews believed and advertised that their respective R&R 14 platforms increased conversion rates. GX90 (Luedtke Dep. 125:3-6, 125:9-13); Trial Tr. 15 349:24-350:10 (Luedtke); GX92 (Osborne Dep. 119:14-120:4, 120:24-121:4); GX482 at - 16 108, -110; GX485 at -122, -134. R&R can increase conversion rates 50 to 150 percent. 17 GX87 (Godfrey CID Dep. 80:11-16, 81:12-13); GX846 at -082; see also 18 (reviewed in camera). 19 B. Product Comparison Between Bazaarvoice and PowerReviews 20 21 16. Before the acquisition, Bazaarvoice and PowerReviews each offered sophisticated R&R 22 platforms to large enterprise manufacturers and retailers. These platforms offered 23 syndication, moderation, analytics and SEO benefits. They differed in their architecture 24 (Bazaarvoice hosts review content on its own servers, while PowerReviews delivered 25 reviews to its customers daily), customization (Bazaarvoice offers a very customizable 26 solution, suitable to large retailers and brands, whereas PowerReviews’s Enterprise 27 platform lacked the same degree of flexibility), and moderation (Bazaarvoice had a large 28 14 1 team of human moderators who read each review, while PowerReviews did not offer 2 human moderators). 3 17. Notwithstanding those differences, the PowerReviews Enterprise R&R platform was a full- 4 featured R&R platform competing directly against Bazaarvoice. GX641; Trial Tr. 347:17- 5 24 (Luedtke). The target market for PowerReviews’s Enterprise R&R platform was 6 companies with over $5 million in annual online sales. GX88 (Hossain Dep. 24:18-25:2, 7 25:5-18); see also GX454 at -335. In comparing its product to in-house solutions, 8 PowerReviews stated that reviews are “a system, not a feature.” GX272 at -218. 9 18. Both companies’ R&R platforms offered syndication of R&R content to retailers and manufacturers. BrandShare, the PowerReviews syndication service, allowed 11 United States District Court Northern District of California 10 PowerReviews retail customers to receive syndicated reviews from their manufacturing 12 partners. GX90 (Luedtke Dep. 57:1-12); Trial Tr. 355:13-25 (Luedtke). With its stronger 13 relationships with manufacturers, Bazaarvoice maintained a syndication advantage over 14 PowerReviews. Its syndication volumes have been increasing “30-50% year over year.” 15 Trial Tr. 889:25-890:15 (Collins). 16 19. In addition to its Enterprise R&R platform, PowerReviews also offered a turn-key R&R 17 solution (ready for immediate use without customization) called PowerReviews Express. 18 PowerReviews Express lacked the features and functionality available in the enterprise 19 R&R platforms offered by Bazaarvoice and PowerReviews and was a low cost, self- 20 service solution for small and medium-sized internet retailers who have different needs 21 than larger companies. Customers purchased PowerReviews Express directly from the 22 PowerReviews website or their eCommerce platform provider’s app store, often using a 23 credit card—the monthly license fee for PowerReviews Express never exceeded $500 per 24 month. See generally GX232 at -767; GX90 (Luedtke Dep. 44:12-45:1). 25 26 20. PowerReviews Express did not compete with either the PowerReviews Enterprise platform or the Bazaarvoice platform. 27 28 15 1 C. Selling R&R Platforms 2 21. Bazaarvoice charges its customers a subscription-based fee for the services it provides. 3 These fees can total hundreds of thousands of dollars annually. GX203 at -104-107. 4 Several years after it launched, PowerReviews changed its business model from one based 5 on free services supported by advertising to a model based on charging subscription-based 6 license fees, like Bazaarvoice. PowerReviews experienced rapid growth after switching its 7 business model. 8 9 22. Enterprise-level R&R platforms are sold in a direct sales process that requires a significant amount of time and negotiation. Prices are negotiated in light of each customer’s demands, taking into account competitive alternatives. Bazaarvoice calls this method of 11 United States District Court Northern District of California 10 setting prices “value-based” pricing, meaning “[t]he more value the [customer] perceives, 12 the higher [Bazaarvoice’s] price point.” GX87 (Godfrey Dep. 103:10-16). 13 23. To develop a pricing proposal, Bazaarvoice collects detailed information from a customer, 14 including total online sales, number of online transactions, and expected sales growth. 15 During the sales process, it is typical for a salesperson to ask the prospective customer to 16 divulge detailed information related to its business, which may include information related 17 to (1) annual volume of online sales; (2) product return rates; (3) historic conversion rates; 18 (4) eCommerce vendor relationships; or (5) project budgets. Bazaarvoice sales 19 representatives will also seek information from customers to determine the customer’s best 20 alternative to purchasing a R&R platform from Bazaarvoice. Both Bazaarvoice and 21 PowerReviews also took into account a customer’s average order volume, average product 22 price, and the customer’s R&R budget when developing pricing proposals. 23 24. After acquiring the relevant information about the prospect, the R&R platform provider 24 offers a price based on its perception of what the prospect is willing to pay. This price is 25 derived from a Return on Investment (“ROI”) calculator. Both Bazaarvoice and 26 PowerReviews used an ROI calculator to price their products based on the ROI it projected 27 prospective customers would achieve using R&R. In light of Bazaarvoice’s value-based 28 16 1 pricing method, Bazaarvoice’s costs are not “a meaningful consideration in how 2 [Bazaarvoice] price[s].” GX81 (Collins Dep. 119:5-21). 3 25. Once a value-based price is calculated, “[o]ther factors such as market competition and 4 budgetary constraints may reduce [Bazaarvoice’s] pricing range.” GX203 at -105. The 5 Bazaarvoice sales force is trained to ask whether a prospective customer is considering 6 other competitive alternatives. The Bazaarvoice sales force frequently learned what prices 7 PowerReviews was offering to customers. See, e.g., GX92 (Osborne Dep. 173:25-174:8); 8 Trial Tr. 706:12-707:20 (Osborne); GX1119 (Godfrey Dep. 19:11-14); GX193 (Walgreens 9 Dep. 54:2-5); Trial Tr. 299:2-12 (Levin/Clorox). In most cases, the presence of competition is relatively transparent. GX494 (Godfrey 30(b)(6) Dep. 69:5-9, 69:11-12). A 11 United States District Court Northern District of California 10 R&R platform vendor seeking to unseat an incumbent R&R platform provider used by a 12 prospective customer can identify the incumbent by viewing the source code of the 13 company’s website, even if the customer did not reveal the incumbent during negotiations. 14 GX90 (Luedtke Dep. 259:20-260:4); GX130 (Dick’s Sporting Goods Dep. 45:1-7, 45:10- 15 12). 16 26. Bazaarvoice and PowerReviews have also been able to use information obtained during the 17 sales process to determine whether an in-house R&R solution is an economically viable 18 alternative for a particular customer. GX781 at -631-645; GX627 at -997; see also 19 GX365*; GX279* at -120; GX807 at -791. In deciding whether to implement a particular 20 solution, customers consider a variety of factors, including acquisition cost, 21 implementation cost, financial viability of the vendor, impact on the eCommerce website’s 22 performance, anticipated ROI, delay until ROI is realized, cost and likelihood of future 23 updates, compatibility, and whether the R&R solution will be manageable for the in-house 24 technology team. Bazaarvoice contended that customers control information shared in 25 negotiations with vendors of R&R, but it is the vendor who controls the critical (value 26 pricing) information. 27 28 27. Prospective customers routinely reveal the identity of competitors during negotiations and may even reveal the terms of competitive offers to improve their bargaining positions. See, 17 1 e.g., GX92 (Osborne Dep. 172:11-21, 172:24-173:6); GX77 (Heverley/BFG Dep. 64:12- 2 16); see also Trial Tr. 456:8-457:19 (Cunningham/BJ’s Wholesale). Accordingly, 3 Bazaarvoice and PowerReviews would adjust their pricing to account for other competitive 4 offers, depending on the nature of the threat posed by the competition. GX92 (Osborne 5 Dep. 173:25-174:8, 225:21-24, 226:3, 252:22-253:4, 253:8-9); Trial Tr. 709:1-11 6 (Osborne); GX58 at -691-693; GX193 (Walgreens Dep. 52:1-15, 52:25-53:7, 53:21-54:5); 7 GX77 (Heverley/BFG Dep. 65:18-20, 65:22-67:4); GX480 at -293, -295; GX279* at -122. 8 See generally GX200 at -156; GX56 at -815; GX900. For instance, there were 9 circumstances where Bazaarvoice offered to match PowerReviews’s price, offered a lower price, or offered products for free. Trial Tr. 709:1-710:14 (Osborne). Brett Hurt, 11 United States District Court Northern District of California 10 Bazaarvoice’s co-founder, former CEO and current Vice-Chairman of the Board, used an 12 example of value-based pricing to share with potential investors: for clothing retailer , “the value delivered was in the 13 of lift for [year] 1. This drove an 14 initial pricing of , which was discounted down to to win against PowerReviews, 15 at 16 threat, one Bazaarvoice sales person said that “[w]hen competition is strong . . . I probe 17 hard on budget and how we are stacking up. The strategy for these is to build enough 18 value and enough of a relationship so that the client wants BV and . . . tell[s us] what it will 19 take . . . . Negotiation of pricing and terms follows.” GX405* at -003. When there was 20 not a “strong competitive landscape,” however, the sales representative “planted our flag . . 21 . and didn’t move it.” GX405* at -002. .” GX203 at -105; see also GX405* at -002-3. When aware of a competitive 22 28. Since prices are individually negotiated, each customer’s price is independent of the prices 23 other customers pay. GX90 (Luedtke Dep. 91:24-92:8, 129:10-13, 129:23-130:4); Trial 24 Tr. 348:23-349:2, 353:22-25, 355:2-5 (Luedtke); GX840 at -944-945; GX92 (Osborne 25 Dep. 105:23-106:1, 111:8-10, 111:12, 115:11-12); Trial Tr. 704:4-7 (Osborne); GX81 26 (Collins Dep. 69:11-70:5, 76:17-77:16); Trial Tr. 800:13-24 (Collins); see also GX305; 27 GX405*. As a result, customers commonly receive different prices, even when purchasing 28 similar products and services. GX92 (Osborne Dep. 146:19-147:5, 162:12-20); Trial Tr. 18 1 704:4-7 (Osborne); GX361*; GX81 (Collins Dep. 77:24-78:4); Trial Tr. 800:25-801:4 2 (Collins). 3 29. Because of the differences between Bazaarvoice’s and PowerReviews’s solutions, not 4 every company was interested in both products. PowerReviews succeeded with small and 5 medium sized business (“SMB”) and the mid-market. GX86 (Dodd Dep. 40:4-13); Trial 6 Tr. 1114:1-9, 529:17-24. Bazaarvoice identified PowerReviews’s “sweetspot” as “the 7 SMB market,” while positioning itself as the R&R provider “built for enterprises.” 8 DX0624 at BZ-01328389; Trial Tr. 442:6-12 (Luedtke), 830:24-831:5 (Collins). 9 Bazaarvoice understood that PowerReviews “may service more websites,” but emphasized the fact that “the vast majority of their 5,000 clients are SMBs using PowerReviews 11 United States District Court Northern District of California 10 Express.” DX624 at -391. Bazaarvoice traditionally served larger retailers and brands, 12 like Wal-Mart and Best Buy, who were very concerned about their sites and required a lot 13 of customization. Trial Tr. 1109:15-20 (Godfrey). But PowerReviews also served a few 14 large retailers, such as Toys “R” Us and Staples, and roughly 80 retailers listed in the 15 IR500. PowerReviews was not as attractive to manufacturers (also referred to as 16 “brands”). Trial Tr. 229:9-17 (Hurt). The average Bazaarvoice customer paid 17 significantly more than the average PowerReviews customer for the solutions they 18 purchased. 19 D. The Broader Social Commerce and eCommerce Economy 20 21 30. R&R is one of a growing number of social commerce tools that have the ultimate goal of 22 converting website visitors to purchasers. Other tools include customer commentary, 23 blogs, imagery (i.e., pictures), Facebook shares and “likes,” and others. The social 24 commerce tools industry is new, dynamic, and evolving at a fast pace. Trial Tr. 759:18-25 25 (Collins), 1400:21-1401:5 (Meredith). It is a “constantly evolving space” which will 26 continue to change the nature of competition among vendors of social commerce tools, 27 including the tools themselves. DX1432 at 6. Because the social commerce industry is at 28 an early stage of development, rapidly evolving, fragmented, and subject to potential 19 1 disruption by technological innovations, the future composition of the industry as a whole 2 is unpredictable. 3 31. All vendors of social commerce tools, including R&R, compete at some level for the social 4 commerce “wallet” because customers have limited resources to dedicate to a variety of 5 social commerce options. There is pressure on providers of social commerce tools to 6 innovate in order to stay competitive. As an example, although R&R functionality was 7 novel when it was released in 2005, today, some consider basic R&R functionality a 8 commodity. R&R software is more widely available and easier to develop today than it 9 was five years ago. 32. R&R vendors differentiate their products and try to add value to their solution by offering 11 United States District Court Northern District of California 10 different features. They may also differentiate their products by offering customization to 12 suit the particular needs of a customer. Some customers value customization and are 13 willing to pay more for a vendor that offers a highly customizable solution. eCommerce 14 sites make a trade-off between complexity and simplicity in selecting an R&R solution; a 15 complex R&R solution may not be appealing to some consumers for price or other reasons. 16 33. eCommerce platforms provide the infrastructure for an underlying eCommerce website, 17 and are the “host” platform upon which all social commerce features are added. There are 18 several basic categories of webpages which most eCommerce platforms offer. eCommerce 19 platform providers vary in sophistication and price, with some eCommerce vendors 20 targeting larger sites and others targeting smaller sites. Oracle Commerce, IBM 21 WebSphere Commerce, and hybris, which was acquired by SAP in June 2013, are in the 22 “top quadrant” of eCommerce platform providers and offer a wide range of built-in 23 features. DX1432 at 1, 14. 24 34. eCommerce is growing at about 16 percent a year while retail stores are growing at three to 25 four percent a year. Trial Tr. 1460:2-8 (Goldberg); DX1432 at 7-8. eCommerce is 26 becoming an increasingly important part of retailers’ and brands’ overall operations, 27 influencing online and offline sales. As such, brands and retailers are investing to improve 28 the quality of their eCommerce websites, and differentiating themselves from their 20 1 competitors. Trial Tr. 1498:19-1499:17 (Goldberg), 1850:24-1852:16 (Meisels); DX1432 2 at 63. 3 III. WHY BAZAARVOICE PURCHASED POWERREVIEWS 4 5 6 A. PowerReviews Was Bazaarvoice’s Only Real Commercial Competitor 35. While intent is not an element of a Section 7 violation, the admissions made by 7 Bazaarvoice and PowerReviews officers and employees prior to the merger during their 8 intense competition with each other undergirds the government’s case and conflicts with 9 much of Bazaarvoice’s presentation at trial. For that reason, the history that follows 10 matters in the analysis of this case. United States District Court Northern District of California 11 36. Bazaarvoice recognized that the acquisition of PowerReviews would eliminate its primary 12 commercial competitor, allowing it to scoop up customers that it would otherwise have to 13 expend $32-50 million to win over from PowerReviews, raise prices, and discourage any 14 new competitive threats in its existing space while pivoting to a bigger opportunity through 15 its control of UGC in the broader eCommerce market. See GX 1148. In the years 16 preceding the acquisition, Bazaarvoice executives repeatedly expressed that PowerReviews 17 was its most significant, and often only, R&R platform competitor. Hurt described 18 PowerReviews as Bazaarvoice’s “fiercest competitor” (GX417 at -036; GX89 (Hurt Dep. 19 348:3-7)), “only real competitor” (GX416 at -683), and “biggest competitor” (GX418 at - 20 912; GX89 (Hurt Dep. 367:17-369:11)); Trial Tr. 130:7-12 (Hurt). Hurt viewed 21 PowerReviews as Bazaarvoice’s number one competitor that was “challenging 22 [Bazaarvoice’s] price points” and winning “key accounts like REI and Staples.” GX408 at 23 -667; GX417 at -036. 24 37. Former Bazaarvoice Chief Revenue Officer Michael Osborne testified that most of 25 Bazaarvoice’s retail deals “involve consideration of PR [PowerReviews] at some point.” 26 Trial Tr. 708:11-23 (Osborne); GX223. Hurt testified that PowerReviews was 27 Bazaarvoice’s largest threat for retailers and manufacturers. GX89 (Hurt Dep. 219:4-7, 28 219:10-220:8, 372:18-373:8). Draft talking points prepared for Hurt to announce the 21 1 acquisition to Bazaarvoice employees state that “[s]ix years into our respective journeys, 2 we are both respected leaders in our space - No. 1 and No. 2.” GX838 at -391; see also 3 Trial Tr. 509:5-11 (Defossé) (quoting GX37 at 7 (“[PowerReviews] is currently in 80% of 4 our sales pipeline . . . .”)); Trial Tr. 757:3-16 (Collins); cf. GX492 (Pacitti Dep. 151:8-13, 5 160:20-23, 161:1-3). 6 38. Bazaarvoice’s co-founder, Brant Barton, wrote in an email to Facebook on June 1, 2012, that “last Thursday we announced the acquisition of PowerReviews, our primary 8 competitor.” GX507 at -412. Bazaarvoice’s CEO, Stephen Collins, similarly testified that 9 for ratings and reviews in the United States, PowerReviews was “the competitor we looked 10 at as the next best competitor” to Bazaarvoice. GX81 (Collins Dep. 143:7-17). At trial, he 11 United States District Court Northern District of California 7 described PowerReviews as Bazaarvoice’s “primary competitor in retail.” Trial Tr. 12 830:19-23 (Collins). 13 39. PowerReviews executives agreed. They repeatedly acknowledged that Bazaarvoice was its 14 most significant and, often, only competitor. PowerReviews’s CEO at the time of the 15 merger, Ken Comée, testified that Bazaarvoice was PowerReviews’s number one 16 competitor. GX84 (Comée Dep. 96:24-97:4, 97:6-9). Its former Vice President of 17 Marketing, Nadim Hossain, testified that he could only think of one customer in the United 18 States that PowerReviews lost to another competitor besides Bazaarvoice. GX88 (Hossain 19 Dep. 55:13-56:22). 20 40. Retailers and manufacturers also viewed Bazaarvoice and PowerReviews as the two most 21 significant providers of R&R platforms. See, e.g., GX193 (Walgreens Dep. 44:1-6, 44:12- 22 14, 44:20-23, 45:1, 49:9-51:8, 51:12-14, 51:17, 66:10-67:3, 67:6); GX192 (Vitamin 23 Shoppe Dep. 40:13-20, 40:22, 61:2-61:9, 61:21-63:18, 63:21-23, 63:25-64:8, 64:10-21, 24 66:19-21, 66:23-67:8, 67:10-11, 67:16-68:7, 68:9-11, 68:13-69:17, 69:19-70:4, 70:7-11, 25 70:14-16); Trial Tr. 1378:1-7 (Beebe/Vitamin Shoppe); GX77 (Heverley/BFG Dep. 84:23- 26 85:19); GX145 (h.h. gregg Dep. 13:24-14:25, 15:8); GX142 (Green Mountain Coffee Dep. 27 31:23-33:6). Customers often looked exclusively at Bazaarvoice and PowerReviews when 28 searching for a R&R platform. GX193 (Walgreens Dep. 66:10-67:3, 67:6); GX125 22 1 (Clorox Dep. 126:19-22, 127:1-10, 127:12-15, 130:17-21, 130:23-131:1); GX76 2 (Waltzinger/BBBeyond Dep. 96:7-17); Trial Tr. 571:18-25 (Waltzinger/BBBeyond). 3 41. Competition between Bazaarvoice and PowerReviews reached its zenith in the summer of 4 2011. By 2011, PowerReviews had secured more than 1,400 customers. GX610* at 4, 10; 5 GX1110 at -509; see also Trial Tr. 164:12-165:2 (Hurt). It also had at least 80 customers 6 listed in the IR500 and roughly 400 enterprise customers. GX 429 at -909. Its retail 7 customer base was larger than Bazaarvoice’s. 8 9 42. PowerReviews positioned itself as a low-price alternative to Bazaarvoice and set an internal goal to “[b]e in every deal [Bazaarvoice] is in,” and encouraged price competition by building a “cost structure to support price compression.” GX277 at -870; GX490 at - 11 United States District Court Northern District of California 10 727; see also GX88 (Hossain Dep. 82:17-85:7, 95:13-16). Bazaarvoice employees viewed 12 PowerReviews as “an ankle-biter that cause[d] price pressure in deals” and acknowledged 13 that many customers brought PowerReviews into negotiations as a “lever to knock 14 [Bazaarvoice] down on price.” GX477* at -466. Bazaarvoice acknowledged that 15 competition between Bazaarvoice and PowerReviews resulted in lower prices to its 16 customers, saying it was “common for PowerReviews to provide extremely low pricing to 17 our clients to push them in their favor.” GX87 (Godfrey CID Dep. 149:25-150:18); Trial 18 Tr. 762:3-5 (Collins); see also GX90 (Luedtke Dep. 278:2-10); Trial Tr. 372:21-373:5 19 (Luedtke); GX82 (Barton Dep. 254:13-256:7). 20 43. PowerReviews’s aggressive approach to pricing frequently forced Bazaarvoice to defend 21 its more expensive prices. Bazaarvoice’s Vice President of Retail Sales, Paul Dodd, 22 warned in July 2011that PowerReviews had adopted a “scorched earth approach to 23 pricing,” which “force[d] all of [Bazaarvoice’s] current prospects and customers to at least 24 understand how and why there is such a delta in price.” GX208* at -475. If a prospective 25 customer was unwilling to pay a premium over the PowerReviews price, Bazaarvoice 26 frequently matched the PowerReviews price or offered a lower price than PowerReviews. 27 GX293 at -118; GX804 at -084; GX1119 (Godfrey Dep. 65:23-67:10); GX92 (Osborne 28 Dep. 205:9-12, 205:15-22, 205:25); Trial Tr. 709:1-11 (Osborne). 23 1 44. PowerReviews actively sought opportunities to steal Bazaarvoice customers by offering 2 lower pricing or other incentives to those customers to switch to PowerReviews. In the 3 summer of 2011, PowerReviews was “[a]ggressively going after [Bazaarvoice’s] existing 4 customer base” and “[d]isrupting our new client sales efforts” by “[l]owering price floors.” 5 GX212 at 14; see also Trial Tr. 1114:10-1115:9 (Godfrey); GX1119 (Godfrey Dep. 45:2- 6 10, 47:10-19, 48:14-49:6); GX85 (Defossé Dep. 229:2-16); GX422*. Bazaarvoice 7 executive Alan Godfrey commented that PowerReviews was convincing key customers 8 like Best Buy, Wal-Mart, and QVC “to evaluate alternatives, or at least, negotiate us to 9 lower price points.” GX915 at -689. 10 45. In some cases, PowerReviews convinced Bazaarvoice customers to switch platforms. United States District Court Northern District of California 11 GX753 at -574; GX736; GX737; Trial Tr. 373:16-18 (Luedtke). In other cases, an offer 12 from PowerReviews provided additional leverage for the customer to negotiate more 13 favorable terms from Bazaarvoice. GX956. Even if PowerReviews was unable to win a 14 customer’s business, its low prices set the bar for negotiations and compressed 15 Bazaarvoice’s margins. See, e.g., GX386* at -414. 16 46. PowerReviews also threatened Bazaarvoice’s dominance in syndication. Bazaarvoice’s 17 network allows R&R content to be shared, or syndicated, for example, between brand 18 manufacturers and their retailers. As more manufacturers sign up for Bazaarvoice’s R&R 19 platform, the Bazaarvoice network becomes more valuable to retailers because it allows 20 them to gain access to a greater volume of R&R by publishing “syndicated” R&R content 21 received from manufacturers’ websites. Trial Tr. 121:24-122:10, 123:6-16 (Hurt). 22 Similarly, as more retailers purchase Bazaarvoice’s R&R platform, the Bazaarvoice 23 network becomes more valuable for manufacturers because it allows them to syndicate 24 content to a greater number of retail outlets. Trial Tr. 121:24-122:10, 123:6-16 (Hurt). 25 The feedback between the manufacturers and retailers creates a network effect that is a 26 significant competitive advantage for Bazaarvoice. Trial Tr. 357:16-358:5, 423:6-25 27 (Luedtke); 711:4-8 (Osborne); GX1220 at -614. 28 47. PowerReviews announced the Open Social Commerce Network, offering a free one-year 24 1 trial to syndication customers in an attempt to diminish Bazaarvoice’s network advantage 2 in July 2011. GX90 (Luedtke Dep. 213:14-214:1, 214:4-17, 216:16-218:14, 218:16-18, 3 219:3-8); GX412 at -494-96. PowerReviews did not create any new technology in 4 conjunction with this announcement, but the marketing of the offering amplified its threat 5 to Bazaarvoice because it publicized that PowerReviews facilitated syndication between 6 manufacturers that were not PowerReviews customers (including manufacturers that were 7 Bazaarvoice customers) and retailers using the PowerReviews platform. GX916; GX88 8 (Hossain Dep. 148:6-150:2, 150:5-17); Trial Tr. 482:18-483:1, 484:20-485:19 (Defossé); 9 Trial Tr. 393:23-394:4. Collins, then Bazaarvoice’s CFO, suggested to Hurt that Bazaarvoice could either compete against PowerReviews and “crush” them, or “damnit lets 11 United States District Court Northern District of California 10 just buy them now.” GX412 at -495. 12 48. Bazaarvoice recognized that increasing competition from PowerReviews threatened the 13 long-term value of Bazaarvoice’s business. In July, 2011, Dodd prepared an analysis for 14 Bazaarvoice executives to see how Bazaarvoice would be affected as it moved down- 15 market into the mid-market and SMB where PowerReviews had historically been 16 successful. GX86 (Dodd Dep. 59:24-60:04). He concluded that price competition from 17 PowerReviews could result in Bazaarvoice’s losing significant revenues. GX553 at -439; 18 see also GX86 (Dodd Dep. 30:17-23, 33:22-25). 19 49. Dodd’s analysis compared the price Bazaarvoice targeted for its R&R platform, based on 20 Bazaarvoice’s “value-based” pricing, to the price that PowerReviews charged for its R&R 21 platform, based on pricing that prospective Bazaarvoice customers claimed to have 22 received from PowerReviews. GX86 (Dodd Dep. 61:10-62:13, 63:2-15). For example, for 23 large enterprise customers (or “A3 customers”), Bazaarvoice’s targeted price was 24 , compared to PowerReviews’s price of . GX553 at -440; see also GX86 25 (Dodd Dep. 46:19-47:5, 61:10-20, 63:2-8). Matching PowerReviews’s price for large 26 enterprise customers would result in a discount of 27 value-based price. GX553 at -440. For medium-sized customers (“A2 customers”), 28 Bazaarvoice’s targeted value-based price was 25 percent off Bazaarvoice’s estimated , compared to PowerReviews’s price 1 of 2 Matching PowerReviews’s price for medium-sized customers would result in a 3 discount off Bazaarvoice’s value-based price. GX553 at -440. Dodd’s analysis showed 4 that competitive pressure from PowerReviews could result in Bazaarvoice’s losing 5 millions of dollars. GX553 at -437, -439; see also GX718 at -092 (showing that Godfrey 6 replies to email thread regarding Dodd’s analysis titled “Suffocate PR plan” and notes “we 7 are also seeing substantial pressure from PR on the high end”); Trial Tr. 1114:17-1115:9 8 (Godfrey). No definitive action was taken with respect to Dodd’s recommendations on 9 pricing. GX86 (Dodd Dep. 85:7-15). 10 . GX553 at -440; see also GX86 (Dodd Dep. 46:19-47:5, 61:10-20, 63:2-4). percent 50. Bazaarvoice’s sales organization repeatedly focused its sales efforts on competing with United States District Court Northern District of California 11 PowerReviews. Bazaarvoice recognized the pricing pressure that PowerReviews imposed 12 and developed competitive strategies to respond. In response to a PowerReviews 13 campaign targeting Bazaarvoice’s manufacturing customers, Bazaarvoice put into motion a 14 plan to “steal one or more major [PowerReviews] clients . . . by offering them something 15 that they can’t refuse.” GX544 at -797. This strategy was intended to send a signal to 16 PowerReviews that Bazaarvoice was willing “to absorb some pain in return for handing 17 [PowerReviews] major client losses.” GX544 at -797. On some occasions, Bazaarvoice 18 succeeded in persuading PowerReviews customers to switch. See, e.g., GX471; GX358 at 19 -612. 20 51. An example of this competition occurred with BestBuy. In July 2011, PowerReviews 21 convinced BestBuy to reevaluate its relationship with Bazaarvoice. Afterwards, 22 Bazaarvoice’s Chief Strategy Officer, Mike Svatek, expressed concern that Bazaarvoice 23 was “seeing new competitive pressure” from PowerReviews through an “aggressive blitz 24 campaign.” GX541 at -266; GX914 at -275. Svatek believed Bazaarvoice needed to 25 “eradicate” PowerReviews, and he proposed a counterattack on PowerReviews’s base. 26 GX914 at -274. He advocated an “aggressive” approach to “unseat” PowerReviews from 27 three of its largest accounts. GX541 at -265. “Whatever the package, I want to screw PR 28 26 1 and out them on their heels by wrecking a few of their big accounts and getting us a couple 2 supernodes. You in? Let’s crush these MFs.” GX397* at -058. 3 52. Hurt recognized another risk posed by PowerReviews: “PowerReviews, while 4 considerably smaller than us today in terms of revenue, still has a relatively significant 5 reach and from this base they could mount a significant challenge especially if acquired by 6 a well-capitalized player in the social commerce or social media marketing solutions space. 7 Thus, leaving them out there is a significant risk.” GX1175 at -904; see also GX326 at - 8 200; GX81 (Collins Dep. 328:3-329:23, 368:15-369:11, 372:1-7); GX82 (Barton Dep. 9 185:23-188:20, 217:12-218:14); Trial Tr. 650:12-651:4 (Barton); Trial Tr. 755:9-23 10 United States District Court Northern District of California 11 (Collins). 53. Bazaarvoice launched an initiative to respond to heightened competition from 12 PowerReviews called “Project Menlogeddon,” a name inspired by PowerReviews’s 13 primary financial backer, Menlo Ventures. GX1084 at -022. Collins described Project 14 Menlogeddon as “a special project to defeat our only meaningful competitor . . . .” GX334 at 15 -684; GX81 (Collins Dep. 417:24-418:5, 418:14-25); see also GX37; GX39. 16 54. During Project Menlogeddon, Bazaarvoice sought to fend off the PowerReviews assault by 17 “building moats” around its most significant customers and enticing large PowerReviews 18 customers to switch to Bazaarvoice’s R&R platform. GX33. As Bazaarvoice’s Chief 19 Marketing Officer, Erin Nelson, put it: “Important we show that we’ve got a plan of action 20 that’s broad and bold[;] Take their top customers[;] Build moats around our top clients[;] 21 Take their data[;] Shake their confidence[.]” GX33; see also GX334 at -684. As part of 22 its plan to “[t]ake it to PowerReviews” (GX40 at -033), Bazaarvoice created pricing 23 guidelines to steal “marquee” PowerReviews customers “at all costs.” GX39 at -941; see 24 also Trial Tr. 501:18-22 (Defossé). Dodd reported to his team that he had met with Collins 25 and other members of Bazaarvoice’s Executive Leadership Team and Collins wanted to 26 “squeeze PR at every point” including “dropping the price floor” to win certain key 27 accounts. GX558; GX86 (Dodd Dep. 125:12-128:6); see also GX81 (Collins Dep. 122:3- 28 6, 122:8-11, 155:8-15). 27 1 55. The warlike language used by Bazaarvoice during Project Menlogeddon is striking. For 2 example, Erin Defossé, Bazaarvoice’s Vice President of Strategy, recommended that the 3 mascot for Project Menlogeddon be the AC-130 military ground attack aircraft because 4 “[i]t is badass and frankly very intimidating.” GX716 at -961. In a “Menlogeddon 5 Update” email, he wrote, “The BV battleship (or AC-130 gunship, rather) and its guns 6 have kicked in and lead rain is starting to drop on PR . . . .” The update detailed both 7 defensive and offensive strategies targeting PowerReviews. GX37 at -947; see also GX35; 8 GX36; GX755. 56. PowerReviews shifted to a “Blue Ocean” business strategy in late fall of 2011—seeking to 10 capitalize on meeting unmet demand in a new market (Trial Tr. 386:13-18 (Luedtke))—to 11 United States District Court Northern District of California 9 enable it to compete less directly with Bazaarvoice because it had been unsuccessful 12 targeting Bazaarvoice’s enterprise clients. 6 At trial, Bazaarvoice argued that this change in 13 strategy meant that PowerReviews was an ineffective competitor that had to stop 14 competing directly with Bazaarvoice prior to the acquisition. But there is no doubt that up 15 to the time of the merger, PowerReviews continued to compete with Bazaarvoice. Trial 16 Tr. 424:21-425:2 (Luedtke); Trial Tr. 1804:14-21, 1825:3-4, 1826:22-1827:5, 1828:17-22, 17 1829:12-1830:13, 1837:10-1838:11 (Comée); Trial Tr. 758:17-20 (Collins). 57. In fact, leading up to the merger, a member of the PowerReviews Board pushed its CEO to 18 19 “push hard on BV flips with very aggressive pricing.” Trial Tr. 1827:9-1828:3 (Comée). 20 Bazaarvoice’s CEO acknowledged that “prior to the merger, PowerReviews put pricing 21 pressure on Bazaarvoice.” Trial Tr. 762:3-5 (Collins). At the time of the merger, 22 PowerReviews executives pitched the merger as one that would result in a “monopoly in 23 the market.” Trial Tr. 1819:11-14, 1820:16-1821:1 (Comée). 58. Further, PowerReviews contemplated that R&R platforms would remain the foundation of 24 25 its offering while employing the Blue Ocean strategy. Trial Tr. 1845:9-12 (Comée). 26 Despite telling Bazaarvoice that his plan was to “pivot the company and focus 27 6 28 In contrast, the “Red Ocean” refers to blood in the water caused by extensive head-to-head competition. 28 1 downmarket,” Trial Tr.1833:20-25 (Comée), PowerReviews’s CEO issued a press release 2 to “send a message to Bazaarvoice,” Trial Tr. 1834:1-3 (Comée), and “emphasize[] the 3 company’s success with large clients,” Trial Tr. 1828:21-23 (Comée). This continued 4 competition was confirmed post-merger, when Bazaarvoice and PowerReviews exchanged 5 lists of each company’s sales pipeline. Those lists demonstrate that the companies had 6 been pursuing many of the same customers up to the time of the transaction. See GX817*. 7 Comée was confident that he could grow PowerReviews into a big and profitable 8 company. Hossain thought that PowerReviews had the ability to surpass Bazaarvoice to 9 become number one. GX88 (Hossain Dep. 121:25-122:2). 10 United States District Court Northern District of California 11 59. A former PowerReviews employee who is now Vice President of Sales for North America for Reevoo, another provider of R&R, summed up the merger this way: 12 13 GX952*. 14 15 16 17 B. Bazaarvoice’s Pre-Acquisition Rationales For The Merger Were Anticompetitive 60. Bazaarvoice unabashedly used anticompetitive rationales for purchasing PowerReviews 18 throughout the year-long negotiations, as did PowerReviews. When outlining the benefits 19 of the acquisition, Bazaarvoice’s Barton noted that acquiring PowerReviews would 20 “[e]liminat[e] [Bazaarvoice’s] primary competitor” and provide “relief from . . . price 21 erosion.” GX514 at -810; see also GX218 at -594; GX82 (Barton Dep. 94:2-24, 129:10- 22 130:9); Trial Tr. 646:11-647:10 (Barton). He later wrote that the transaction would enable 23 the combined company to “avoid margin erosion” caused by “tactical ‘knife-fighting’ over 24 competitive deals.” GX519 at -751; see also GX82 (Barton Dep. 239:14-240:20, 240:23- 25 242:14). Hurt saw the acquisition of PowerReviews as an opportunity to “tak[e] out 26 [Bazaarvoice’s] only competitor, who . . . suppress[ed] [Bazaarvoice] price points . . . by as 27 much as 15% . . . .” GX518 at -475. He later observed that Bazaarvoice would have “[n]o 28 29 1 meaningful direct competitor” after acquiring PowerReviews, thereby reducing “pricing 2 dilution.” GX899 at -817-818; see also GX1181 at -158; Trial Tr. 259:20-260:13 (Hurt). 3 61. Collins wrote that Bazaarvoice had “[l]iterally, no other competitors,” and he expected 4 “[p]ricing accretion due to [the] combination” of the two firms. GX522 at -035. He 5 pointed out that the acquisition of PowerReviews would “eliminate the cost in time and 6 money to take [PowerReviews’s] accounts” and “reduce [Bazaarvoice’s] risk of account 7 losses as [PowerReviews] compete[d] for survival.” GX324 at -922. The Bazaarvoice 8 “executive team’s views of the transaction’s ability to improve Bazaarvoice’s pricing 9 power” did not change prior to the acquisition of PowerReviews. GX492 (Pacitti Dep. 10 United States District Court Northern District of California 11 168:13-18). 62. PowerReviews’s CEO until March 2012, Pehr Luedtke, also believed the benefits of the 12 merger between Bazaarvoice and PowerReviews included “margin expansion” (GX249 at - 13 940), the “possibility of reducing the discounting that we had seen in the marketplace,” and 14 the fact that it would “[e]liminate[e] competitive risk and an acquisition of PowerReviews 15 by a larger player . . . .” GX253; see also GX90 (Luedtke Dep. 289:10-15, 289:18-24); 16 Trial Tr. 329:2-15, 332:7-333:5 (Luedtke). 17 63. The parties engaged in three separate rounds of negotiations leading up to the merger: (1) 18 during April to May 2011, (2) during November to December of 2011, and (3) during 19 April to May 2012. Trial Tr. 174:8-176:11 (Hurt). In the first round, on April 21, 2011, 20 Barton sent an email to other Bazaarvoice executives regarding why the company should 21 consider acquiring PowerReviews, saying that “taking out one of your biggest competitors 22 can be game-changing.” GX514 at -810; Trial Tr. 645:21-646:10 (Barton). He listed the 23 “Pros” of the deal, including “[e]limination of our primary competitor” and “relief from the 24 price erosion that Sales experiences in 30-40% of deals . . . of up to 15-30%.” GX514 at - 25 810; see also GX82 (Barton Dep. 202:10-203:16); Trial Tr. 646:11-647:10 (Barton). He 26 recognized “the market will place a premium on us having such a dominant market 27 position, which is a powerful competitive moat.” GX514 at -810. He anticipated that “we 28 will be able to retain an extremely high percentage of PR’s customers due to scarce/low 30 1 quality alternatives.” GX514 at -810. He reported that Hurt, Collins, Osborne and Svatek 2 were all “supportive.” GX220 at -112; GX219 at -258. Osborne agreed, writing in an 3 email to Barton on April 24, 2011 that “[i]f we buy them [PowerReviews] it changes 4 everything for our model . . . . Because 10-20% price erosion will disappear . . . . Because 5 this is competitively HUGE . . . .” GX221 at -528; Trial Tr. 718:7-25 (Osborne). 6 64. On May 4, 2011, in an email to Bazaarvoice’s board of directors, Hurt informed the board of an upcoming meeting with senior PowerReviews executives to explore a potential 8 acquisition of PowerReviews. GX518 at -474-75. He wrote, “[p]otentially taking out our 9 only competitor, who is both suppressing our price points (by as much as 15% according to 10 Osborne) . . . could be a highly strategic move . . . .” GX518 at -475 and Trial Tr. 170:20- 11 United States District Court Northern District of California 7 171:11 (Hurt). In his notes from the meeting, Barton wrote that the transaction would 12 enable the combined company to “avoid margin erosion” caused by “tactical ‘knife- 13 fighting’ over competitive deals.” GX519 at -751; see also GX82 (Barton Dep. 239:14- 14 240:20, 240:23-242:14). Not all Board members agreed. Ed Keller rejected the idea that 15 the acquisition would stem any price erosion. GX0518 at -474. Some outside board 16 members agreed with Keller, including Chris Pacitti who spoke with Hurt about the fact 17 that the acquisition would not alleviate pricing pressure, because there would be a swarm 18 of competitors to fill PowerReviews’s place. Trial Tr. 244:21-247:3 (Hurt). 19 65. In an email, Hurt stated that “[t]here is no doubt that PowerReviews brings our pricing 20 down.” GX0424 at -543. He also noted that there were “[n]o major defections to them (or 21 minor ones that I can think of) at any time recently . . . . Our services, breadth of solutions, 22 global reach, large blue chip client base, and network effect (i.e., the BV Network of 23 content syndication) all help build defensive barriers to entry.” GX0424 at -543. 24 66. PowerReviews also believed the merger would allow Bazaarvoice to raise prices. In May 25 2011, Luedtke wrote to Hurt that one of the obvious benefits of a merger between 26 Bazaarvoice and PowerReviews was “margin expansion.” GX249 at -940. Luedtke 27 testified that one of the ways margins can expand is through the elimination of competitive 28 discounting that occurred due to competition between Bazaarvoice and PowerReviews. 31 1 GX90 (Luedtke Dep. 284:8-285:2, 285:5-16, 285:19-25, 286:12-22, 287:1); Trial Tr. 2 329:8-15 (Luedtke). Luedtke recognized that Bazaarvoice would likely pay a premium to 3 acquire PowerReviews because it would allow Bazaarvoice to eliminate the “knife 4 fighting” between the two companies. GX90 (Luedtke Dep. 289:10-15, 289:18-24); Trial 5 Tr. 329:8-15 (Luedtke). He further noted that “[e]liminating competitive risk and an 6 acquisition of PowerReviews by a larger player is worth another . . . $30-50M.” GX253 at 7 -188. He explained that this meant “we wanted to avoid this period, as Gartner explains it, 8 of intense competition and occasionally price compression . . . .” Trial Tr. 404:10-405:2 9 (Luedtke). 10 67. On May 20, 2011, Bazaarvoice executives represented to the company’s board of directors United States District Court Northern District of California 11 that the acquisition of PowerReviews would “[e]liminate [Bazaarvoice’s] primary 12 competitor” and “reduc[e] comparative pricing pressure.” GX521 at -088; Trial Tr. 13 653:19-654:2 (Barton). Bazaarvoice’s board did not proceed with further negotiations at 14 that time. 15 68. Bazaarvoice continued to explore its interest in PowerReviews in October 2011, when 16 Collins wrote: “[h]ow much longer does it take us to win deals because [PowerReviews is] 17 out there and how much does the price competition impair our long-term value as a 18 company?” GX522 at -035. Collins expected the combined company to receive a higher 19 valuation as a public company because it would have “[l]iterally, no other competitors,” 20 and the combined company’s “share of the IR 500 would move close to 50% in absolute 21 terms and higher in retail sales coverage.” GX522 at -035 and Trial Tr. 754:10-18 22 (Collins). He anticipated “[p]ricing accretion due to combination.” GX522 at -035. He 23 added other reasons: “the reason to merge is not as much about taking out the competitive 24 threat as building scale to take on the greater market opportunity and fend off the bigger 25 long term threats.” GX320 at -026. He also noted that the acquisition would benefit 26 customers because it would create “maximum market efficiency particularly in advertising 27 where audience reach at scale is what the advertiser wants from the network owner. ditto 28 on data network.” Id. 32 1 69. In November 2011, Hurt suggested a “second round” of discussions with PowerReviews to 2 the Board because “[c]ombined, we would be approaching the 50% share point of the IR 3 500. There is no other US competitor with more than 10 clients and 104 [of the IR500] do 4 not have any solution or are using an in-house solution. The cost in time and money to 5 displace PowerReviews from these clients is surely a very big number. . . . Here is a list of 6 other potential benefits from a combination: 7 8 9 10 United States District Court Northern District of California 11 12 13 14 1. No meaningful direct competitor. Shortened sales cycles, less pricing dilution. (…) 2. Dramatic increase in reach and overall market share making future competition extremely difficult and will increase switching costs. . . . (…) 7. Help us to focus 100% of our strategy without the whipsaw effect of reacting to a competitor that will only be focused on disrupting our lead position . . . .” 15 GX1175 at -904 and Trial Tr. 175:22-176:11 (Hurt). While Hurt testified that this message 16 was primarily written by Collins, he agreed with many of those points. Trial Tr. 192:2317 193:8, 199:12-200:1, 200:22-201:1 (Hurt). 18 70. On December 9, 2011, Bazaarvoice held a six hour meeting with PowerReviews. 19 Afterwards, Collins wrote a memorandum in which he summarized the benefits of the deal, 20 including the ability to “eliminat[e] feature driven one-upmanship and tactical competition 21 on retail.” GX324 at -921. He also noted the acquisition would “[c]reate[] significant 22 competitive barriers to entry and protect[] [Bazaarvoice’s] flank.” GX324 at -922. As a 23 result of the transaction, he believed the combined company would have “[c]lear market 24 leadership with almost 50% of the IR 500 and influence over $80BB in US online retail 25 sales or nearly half of all U.S. online retail sales.” GX324 at -922. Pacitti passed along 26 these notes to his partners at Bazaarvoice-investor Austin Ventures, adding that 27 “[PowerReviews is] the only other credible player in the space. Buying them removes any 28 33 1 cheap entry point for a future competitor . . . . It improves our pricing power (they are the 2 low cost alternative) and gives us complete ownership of the category and data asset.” 3 GX1181 at -158. 4 71. At that time, Bazaarvoice placed merger discussions on hold in order to conduct its initial 5 public offering (“IPO”). GX954; GX328* at -268; Trial Tr. 1393:12-25 (Meredith). 6 Bazaarvoice filed its final amendment to its S-1 registration statement with the United 7 States Securities and Exchange Commission on February 9, 2012, one day before it began 8 a series of IPO roadshow presentations for potential investors. GX964. The S-1 identified 9 two firms by name in its discussion of “direct and indirect” competitors: PowerReviews and Revieworld (the parent company of Reevoo, Bazaarvoice’s European competitor). 11 United States District Court Northern District of California 10 GX964 at 11. 12 72. PowerReviews executives recognized that Bazaarvoice would likely have a higher public 13 market valuation if it acquired PowerReviews prior to the IPO, as the merger would 14 eliminate Bazaarvoice’s closest competitor. GX90 (Luedtke Dep. 292:22-293:17, 293:20- 15 294:1); Trial Tr. 329:22-25 (Luedtke). Bazaarvoice’s board, however, concluded the 16 acquisition could wait. Bazaarvoice issued its IPO on February 23, 2012, raising $114 17 million. GX805 at -097-98. 18 73. In April 2012, Bazaarvoice entered its final round of negotiations to acquire 19 PowerReviews. PowerReviews executives prepared a presentation in which they described 20 the benefits of a merger with Bazaarvoice. Those benefits included: 21 • “Monopoly in the market”; 22 • “More predictable revenue growth (less competition)”; 23 • “Better monetization w/o pricing pressure”; • “Not having us in the hands of a competitor, especially one with deep pockets and existing channels”; • “Removing competitor clears up [Bazaarvoice’s] story to customers, analysts, employees.” 24 25 26 27 28 GX610* at 27-28. 34 1 74. A PowerPoint presentation at a Bazaarvoice board meeting on April 4, 2012 listed the 2 “primary benefits” of the potential acquisition. It included “[g]reater consumer reach,” 3 “[e]xpands network and market share,” “SMB go-to-market capability,” “[a]llows us to 4 concentrate more resources on Brand business sooner” and “mitigates risk of acquisition 5 by another potential competitor with significant assets to invest in competing.” 6 908; Trial Tr. 1395:8-1396:21 (Meredith). GX429 - 75. Bazaarvoice executives prepared a due diligence memorandum dated May 16, 2012, 7 8 recommending that the company finally acquire PowerReviews. The memorandum noted, 9 in furtherance of Bazaarvoice’s goal to “becom[e] the de facto solution for retailers to capture and use digital word-of-mouth across online, mobile, and in-store channels,” that 11 United States District Court Northern District of California 10 the transaction: 12 accelerates this strategy in two ways. First, [PowerReviews’s] customer base includes 86 IR 500 retailers who have resisted becoming Bazaarvoice customers despite significant attempts to displace [PowerReviews] from these accounts. It is unlikely that we can attract these retailers to our platform in the foreseeable future nor without significant cost. Estimated costs to acquire the [PowerReviews] customer base is $32-50m with a substantial percentage of that being attributed to displacing them in large accounts . . . Further, by addressing the mid-market, Bazaarvoice blocks market entry by competitors and therefore we “cover our flank” to ensure our retail business is protected from direct competition and premature price erosion . . . The addition of major retail hubs such as Staples and Toys R Us strengthens our value proposition to brands and further increases the switching costs, and therefore deepens our protective moat, for brands and retailers alike. 13 14 15 16 17 18 19 20 21 GX925 at -941, -943 7; see also GX93 (Svatek Dep. 252:16-253:9); GX87 (Godfrey CID 22 Dep. 22:3-24:24, 25:2-27:7, 28:7-29:3, 29:23-31:6). 23 76. On May 23, 2012, the Bazaarvoice board discussed the acquisition of PowerReviews. The 24 presentation emphasized the “[e]nhanced value of network economics” and that “[a] 25 combined entity would command 49% of the Internet Retailer 500 . . . [t]he other 26 remaining IR 500 generally have an internal solution or no solution at all . . . . A combined 27 28 7 Another copy of this document, GX1148, was also used during the trial. 35 1 entity would influence 52% of the U.S. online retail sales . . . .” GX332 at -287, -289. It 2 also identified as likely benefits arising from the acquisition “Foregone Customer 3 Acquisition Costs,” including the “cost premium to achieve competitive steals which if, on 4 average, was 25% to 50% equates to an aggregate cost of $40 to $50 million.” GX332 at - 5 292. When explaining the PowerReviews acquisition to Bazaarvoice employees, 6 Bazaarvoice executives noted that it would take years and significant resources to continue 7 competing with PowerReviews for enterprise customers. GX838 at -392. 8 77. On May 27, 2012, Collins sent an email to various employees asking them to discuss with 9 PowerReviews’s Luedtke “ways we can very very quickly market Connections (including BrandVoice) [Bazaarvoice’s syndication feature] to PowerReviews clients . . . .” GX5. 11 United States District Court Northern District of California 10 Collins noted that “[m]any of [PowerReviews’s] clients may have been hearing a 12 ‘syndicate for free’ message, so we’ll have to navigate that,” indicating that Bazaarvoice 13 wanted to charge PowerReviews customers for syndication post-merger which they 14 currently obtained for free. GX5. 15 78. Bazaarvoice acquired PowerReviews on June 12, 2012. The purchase price for the 16 transaction, including cash and non-cash consideration, was valued between $150.8 million 17 and $168.2 million. DX1801 at 4; GX318 (Def.’s Answer ¶ 10). Hurt testified that it was 18 a “very rich deal” and a “very high amount to pay.” GX89 (Hurt Dep. 449:8-25); Trial Tr. 19 250:20-251:2 (Hurt). Bazaarvoice also was concerned about the price and the “message it 20 will send to the Street about BV overpaying, desperation, fearing PR who is 10x smaller.” 21 GX1260 at -826; Trial Tr. 1822:19-1823:10 (Comée). 22 79. Barton testified that “by acquiring PowerReviews we would be acquiring the company 23 that we most often competed with in the U.S. for ratings and reviews.” GX82 (Barton 24 Dep. 202:10-203:16); Trial Tr. 633:10-15 (Barton). As a result of the merger, Bazaarvoice 25 was able to acquire every customer that it had lost to PowerReviews. GX222 at -679. 26 C. Bazaarvoice’s Trial Rationale for the Acquisition 27 80. As described above, anticompetitive rationales infused virtually every pre-acquisition 28 document describing the benefits of purchasing PowerReviews. But there was an 36 1 additional, not inconsistent purpose for the merger. Bazaarvoice pointed to a relatively 2 small number of pre-acquisition documents and the trial testimony of its board members to 3 argue that outside board members saw that Bazaarvoice’s “core business (content capture 4 & display, to put it simply) is trending towards commoditization.” DX0574 at -168. It 5 said the purpose of the acquisition to expand its offerings beyond R&R to: “new initiatives 6 – CI, Network, Advertising – to completely replace [the] core business revenue AND 7 much more.” Collins posited that if Bazaarvoice doesn’t “figure out how to completely 8 replace [the] ‘legacy’ business, someone else will beat [them] to it.” DX574 -168; Trial 9 Tr. 835:6-836:9 (Collins). Bazaarvoice asserted that the ultimate rationale for the acquisition was based upon the belief that Bazaarvoice’s “future solutions are dependent 11 United States District Court Northern District of California 10 on building the largest network possible with the greatest consumer reach.” GX1175 at - 12 904; Trial Tr. 247:15-249:16. 13 81. Collins said that the “most attractive strategy for [Bazaarvoice] . . . is to build a large 14 audience that has lots of content data and then provide tools to allow the retailers, [its] 15 retail clients, to make money from their assets, facilitated by the efficiencies [Bazaarvoice] 16 bring[s] through [its] technology.” Trial Tr. 836:21-839:18 (Collins); DX735. By 17 expanding its number of clients through the acquisition, including retailers and brands, 18 Bazaarvoice is also able to obtain more data variety and volume, and provide better data 19 analytics products. Trial Tr. 1391:6-1392:5 (Meredith). “The more data [Bazaarvoice has] 20 and control[s] the more valuable the company will be regardless of revenue and 21 profitability.” DX735 at -985; Trial Tr. 1391:6-1392:5 (Hurt). 22 82. Collins hopes that the combination of the PowerReviews acquisition and the media 23 advertising strategy embodied in Bazaarvoice’s acquisition of Longboard Media in the fall 24 2012 will permit Bazaarvoice someday to give retailers “a check . . . by engaging in benefit 25 26 27 28 37 1 or site monetization activities enabled by [its] network.” Trial Tr. 839:19-840:6 (Collins). 8 2 The idea is that retailers would be able to make a profit from Bazaarvoice’s strategy of 3 using the larger retail audience and its data bank to sell brands more social commerce 4 functionality, including syndication, data analytics, and advertising features. DX735 at - 5 985. 83. The purchase of PowerReviews furthered those goals. Bazaarvoice acquired between 6 7 2,000 and 4,000 customer relationships, as well as millions of pieces of content that would 8 benefit its customers. Trial Tr. 687:4-687:13 (Barton). The acquisition of PowerReviews 9 amplifies its access to raw UGC and consumer behavior data and “bring[s] significant opportunities for syndication, advertising, and data.” DX21 at 2; Trial Tr. 243:12-244:20, 11 United States District Court Northern District of California 10 248:17-249:16 (Hurt), 674:23-676:5 (Barton), 804:25-806:4 (Collins), 1390:25-1391:5 12 (Meredith); GX419 at -484. It increases the number of potential syndication connections 13 between retailer and brands, which would maximize value for Bazaarvoice’s customers. 14 Trial Tr. 169:17-170:5 (Hurt), 848:10-849:12 (Collins). A larger syndication network also 15 increases the value proposition for brands. “The more content [Bazaarvoice could] publish 16 for [brands], the more consumers that see their reviews, the more value they get, and the 17 more attractive and valuable [the Bazaarvoice] solution [is] for them.” Trial Tr. 840:7- 18 842:1 (Collins); DX649 at -284. 84. The acquisition provided Bazaarvoice with an increased ability to provide current 19 20 Bazaarvoice and PowerReviews customers with more value-added products than basic 21 R&R. Trial Tr. 243:23-244:20 (Hurt). It also allowed Bazaarvoice “to drive advertising 22 value,” because “[a]dvertisers value reach at scale.” DX476 at -083; Trial Tr. 804:25- 23 806:4, 848:10-849:12 (Collins), 1390:25-1392:5, 1429:5-1430:19 (Meredith). Bazaarvoice 24 could increase the number of brand and retail websites to which Bazaarvoice could sell 25 26 27 28 8 Following the acquisition, Bazaarvoice acquired Longboard Media in 2012. Longboard Media is an advertising network that allows brand advertisers to buy advertising inventory on large retail sites where they can influence purchase decisions of shoppers who are on those retail sites. Trial Tr. 656:17-657:6. 38 1 personalized and re-targeted advertising. DX476 at -086; Trial Tr. 848:10-849:12 2 (Collins), 1390:25-1392:5, 1400-2-20 (Meredith). Bazaarvoice believes that it is at “the 3 intersection” of “the space between the cloud, social media and big data” and believes that 4 the acquisition will help develop that space. Trial Tr. 1390:17-1391:5 (Meredith). 5 85. Further, by acquiring PowerReviews Express, Bazaarvoice gained access to 6 PowerReviews’s technology and solutions to sell to SMB sites. DX15 at -245; DX488 at - 7 501; Trial Tr. 775:23-776:11, 804:25-806:4 (Collins). PowerReviews’s solution is useful 8 to Bazaarvoice from a reseller standpoint because it was more self-service and easier to 9 install. Trial Tr. 1808:6-1809:16 (Comée). 86. Downplaying the anticompetitive aspect of the acquisition, during an analyst call prior to 11 United States District Court Northern District of California 10 the acquisition on May 24, 2012, Collins said that pricing considerations were “not the 12 reason for the merger.” Trial Tr. 846:7-849:12 (Collins); GX840 at -938. In an email in 13 late April 2012 and at trial, he also stated that “analysts might expect pricing impacts” as a 14 result of the acquisition, but noted that “this is not our strategy or the reason for the deal.” 15 He continued, “pricing per client may actually come down for retail” after the acquisition. 16 DX1855 at -463; Trial Tr. 843:23-845:12 (Collins). A pursuit of marginal price increases 17 for customers, he argued, would run counter to Bazaarvoice’s long-term strategy of 18 expanding its network to sell advertising to brands. Trial Tr. 849:13-850:14 (Colllins). If 19 Bazaarvoice raised prices, forcing retailers and brands to abandon Bazaarvoice, it would 20 decrease the value of its network, destroy its ability to pursue its new business lines, and 21 stem its long-term growth. Trial Tr. 842:21-843:16 (Collins). Similarly, Hurt testified that 22 alleviating pricing pressure was simply not “a winning argument for the board.” Trial Tr. 23 244:21-247:3, 259:12-19 (Hurt); GX518. 24 87. A critical asset in building a successful social commerce network is to have the largest 25 audience possible because that is how “advertisers and marketers and brands think about 26 the value they get.” Trial Tr. 842:21-843:16 (Collins). In Collins’s view, “any tactical 27 action that would reduce the size of our audience runs counter to the current and long-term 28 strategy for the company, which is to provide brands the maximum access or the largest 39 1 audience they can access with regard to shoppers globally.” Trial Tr. 842:21-843:16 2 (Collins). As a result, “any financial gains that we might obtain in the short-term from 3 raising price would be transitory and could result in losing significantly more money on 4 the brand side because that retailer decides to explore an alternative, whatever that 5 alternative might be.” Trial Tr. 842:21-843:16 (Collins). 6 88. Bazaarvoice asserted that it has maintained a consistent strategy with respect to pricing 7 prior to and after the merger so that it would achieve the long-term objectives of the 8 acquisition. If necessary, Bazaarvoice would “give away or even pay retailers to use [its] 9 platform so [Bazaarvoice could] maximize reach.” DX735 at -985; Trial Tr. 837:21839:18 (Collins). In December 2012, six months after acquiring PowerReviews, Collins 11 United States District Court Northern District of California 10 advised his executive team that the company’s board supports a “no excuses win deals 12 mandate to materially drive up new client adds” – including offering prices that are “client 13 friendly” and without worrying about “bookings dollars as the pass/fail measure.” DX753 14 at -799. 15 89. The long-term future for Bazaarvoice probably is what Collins predicted at trial. But 16 Bazaarvoice’s efforts at trial to walk away from its central rationale leading up to the 17 merger—that acquiring PowerReviews would significantly diminish price competition for 18 R&R platforms—was, at best, unconvincing. Some outside directors may have been 19 skeptical of the anticompetitive rationale at times during the negotiation, but management 20 clearly was not. The lack of pre-acquisition documentary evidence supporting 21 Bazaarvoice’s argument at trial that the elimination of its primary competitor had nothing 22 to do with anticompetitive intent and likely effect, particularly compared with the 23 substantial evidence of anticompetitive intent and likely effect to the contrary, was striking. 24 The anticompetitive rationale for the acquisition remained constant in board presentations 25 throughout the negotiation process, and it underlies many of the documents and much of 26 the testimony that Bazaarvoice relied on at trial. 27 90. Barton’s explanation at trial seemed closest to the mark among the Bazaarvoice witnesses: 28 40 “And I felt the acquisition of PowerReviews would--you know, number one, it would be meaningful and it would give us a larger base of customers to go sell these new products into; and I also felt that it would just buy us some time. You know, I thought of it in terms of if we acquire this company, if we secure a position in the ratings and reviews business and have a large amount of customers, hopefully--I was hopeful that that would discourage a new competitive threat immediately going after that business because I wanted us to have the time and the resources to focus on the bigger opportunity ahead.” Trial Tr. 678:13-22 (Barton). 1 2 3 4 5 91. Bazaarvoice’s post-merger market share, its significantly expanded network and the added 7 content provided by the merger vastly strengthens Bazaarvoice’s R&R platforms against 8 any competitor. The most plausible conclusion from the evidence presented at trial is that 9 Bazaarvoice wanted to buy PowerReviews to use its enhanced market power in R&R 10 platforms to keep competitors out of the R&R market and use its dominance to avoid 11 United States District Court Northern District of California 6 competition in pricing and innovation while also pursuing the other long-term objectives 12 that Collins described. 13 IV. THE RELEVANT MARKET 14 A. The Relevant Product Market is R&R Platforms 15 i. 16 R&R is Table Stakes 17 92. Section 7 of the Clayton Act is primarily concerned with the likelihood of anticompetitive 18 effects. In applying Section 7, courts have traditionally assessed what the relevant product 19 and geographic markets are. See, e.g., Marine Bancorporation, Inc., 418 U.S. at 618 20 (“Determination of the relevant product and geographic markets is “a necessary predicate” 21 to deciding whether a merger contravenes the Clayton Act.”) (citation omitted); U.S. DEPT. 22 OF JUSTICE & FED. TRADE COMM’N, HORIZONTAL MERGER GUIDELINES (2010) 23 (“market definition can be informative regarding competitive effects”). These are 24 questions of fact on which the government bears the burden of proof. HDC Med., Inc. v. 25 Minntech Corp., 474 F.3d 543, 547; Oracle Corp., 331 F. Supp. 2d at 1175. The 9 , at 7 26 27 28 9 The 2010 HORIZONTAL MERGER GUIDELINES will be referred to as MERGER GUIDELINES. Prior MERGER GUIDELINES will be referred to with their year of publication, e.g., MERGER GUIDELINES (2003). 41 1 government urged that the relevant market is R&R platforms sold in the United States to 2 retailers and manufacturers. Bazaarvoice proposed a much larger cluster market including 3 R&R and other social commerce tools, like Q&As, blogs, forums and social networks 4 internationally. The Court finds, in light of the facts below and the expert testimony, that 5 the government met its burden of proof to show that the relevant product market is R&R 6 platforms. 7 93. Witness after witness testified to the effect that R&R is considered “critical” to the 8 business of many online retailers and that it is unlikely that most online retailers would 9 ever eliminate R&R entirely, as consumers now expect ratings and reviews to be a part of their online shopping experience. See, e.g., GX302 at 2; GX81 (Collins Dep. 339:3-8); 11 United States District Court Northern District of California 10 Trial Tr. 752:17-20 (Collins); Trial Tr. 1195:7-16 (Friedland/Build.com); GX193 12 (Walgreens Dep. 43:4-6, 43:12-13); GX111 (Bon-Ton Dep. 44:16-45:1); GX141 13 (Golfsmith Dep. 36:12-21, 58:10-19); Trial Tr. 1218:16-22 (Maki/Golfsmith); GX160 14 (Lord & Taylor Dep. 32:18-25, 33:5-7, 33:10-14); GX108* (Black & Decker Dep. 88:22- 15 89:15); GX168 (Onlineshoes.com Dep. 31:21-24, 32:3-9); GX144 (Hayneedle Dep. 47:5- 16 25); GX105 (Belk Dep. 30:8-18); GX196 (World Kitchen Dep. 76:3-22); Trial Tr. 17 1303:11-16 (Bausch/World Kitchen); GX171 (Pacific Sunwear Dep. 48:19-49:7); GX79 18 (Caine Dep. 25:8-17); Trial Tr. 568:13-20 (Waltzinger/BBBeyond); Trial Tr. 1353:19- 19 1354:11 (Hughes/Astral Brands). 20 94. Over the past five years, an increasing number of online retailers have adopted R&R 21 platforms as reported in the IR500. In 2009, over 300 firms in the IR500 did not use R&R 22 platforms. By 2013, only about 100 firms in the IR500 did not use R&R platforms. Trial 23 Tr. 928:14-18 (Shapiro); GX1038. 24 95. An online retailer without R&R is at a competitive disadvantage. GX131 (Dillard’s Dep. 25 35:5-36:4); GX577* at -038; GX76 (Waltzinger/BBBeyond Dep. 38:11-13, 38:16-22); 26 Trial Tr. 568:13-20 (Waltzinger/BBBeyond). As PowerReviews’s former CEO, Ken 27 Comée, explained, “[R&R] was table stakes. You had to have it.” GX84 (Comée Dep. 28 41:5-11, 41:14-42:8). Bazaarvoice described it the same way. GX577 at -038. 42 1 ii. Bazaarvoice and PowerReviews Described R&R as a Distinct Market and Recognized They Were Duopolists in the Market 2 3 4 5 6 7 8 9 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 96. The Court found it persuasive that in the ordinary course of business, Bazaarvoice and PowerReviews recognized that R&R platforms comprise a distinct market. See United States v. H & R Block, Inc., 833 F. Supp. 2d 36, 52 (D.D.C. 2011) (“When determining the relevant product market, courts often pay close attention to the defendants’ ordinary course of business documents.”); see also F.T.C. v. Whole Foods Mkt., Inc., 548 F.3d 1028, 1045 (D.C. Cir. 2008) (Tatel, J., concurring) (“record evidence indicates that the Whole Foods and Wild Oats CEOs both believed that their companies occupied a market separate from the conventional grocery store industry”). For example, in the materials Bazaarvoice regularly sent to its board members, Bazaarvoice only named other R&R platform providers in the slides titled “Competitive Update.” See GX1030* at -018; GX1031* at 661; GX1032* at -844. Bazaarvoice and PowerReviews’s descriptions of market share and market size also reflect the view that they considered R&R platforms to be a distinct market. See, e.g., GX798 at -937; GX1128 at -404; GX524 at -722; Trial Tr. 628:25630:10 (Barton). 97. Bazaarvoice considered the R&R platform market highly concentrated. Before the merger, Defossé said that the number of R&R platform alternatives available to enterprise customers is limited: “My take is that there really isn’t a market for them to understand (as it relates R&R), it is us or PowerReviews and in the game of big clients: we win.” GX540 at -252. In SEC filings prior to the litigation, Bazaarvoice only named PowerReviews and Reevoo, Bazaarvoice’s European competitor, as direct competitors. See GX964 at 11; GX965 at 13; GX967 at 25; GX968 at 31. After the government filed this action, Bazaarvoice identified Pluck, Gigya, and Viewpoints by name as competitors in its public securities filings, but Collins acknowledged that this lawsuit may have influenced the decision to start naming these companies in those filings. GX81 (Collins Dep. 223:2224:1). See GX970 at 8; Trial Tr. 759:12-17 (Collins). 28 43 1 98. Hurt recognized the duopolistic nature of the R&R platform market when remarking that 2 one of the benefits of acquiring PowerReviews would be “[p]otentially taking out our only 3 competitor, who is both suppressing our price points (by as much as 15% according to 4 Osborne) . . . .” GX518 at -475. At trial, Hurt testified that prior to the merger he believed 5 that PowerReviews was Bazaarvoice’s biggest competitor “for the U.S. market and ratings 6 and reviews.” Trial Tr. 130:7-12 (Hurt). PowerReviews executives similarly saw the 7 R&R platform market as a duopoly between Bazaarvoice and PowerReviews. As recently 8 as April 2012, shortly before the transaction, PowerReviews’s Luedtke noted “wide 9 recognition” of the “duopoly” between Bazaarvoice and PowerReviews. GX254 at 3 and Trial Tr. 340:22-342:8 (Luedtke). Luedtke repeatedly recognized the absence of other 11 United States District Court Northern District of California 10 meaningful competitors, calling the R&R platform market a “duopoly” on several other 12 occasions. See GX255 at -698; GX955 at -568; GX275 at -093; GX489 at -172; Trial Tr. 13 345:24-346:14 (Luedtke). At trial, Luedtke said that he was suggesting only that 14 PowerReviews offered comparable services to Bazaarvoice and that he used “duopoly” to 15 describe that there were two leading providers of outsourced R&R and to get better brand 16 recognition. Trial Tr. 376:14-377:1 (Luedtke). 17 99. PowerReviews’s head of marketing, Cathy Halligan, also saw that PowerReviews was in a 18 “duopoly” with Bazaarvoice. GX683 at -339; GX776 at -766. Halligan wrote that it was 19 well-known in the industry that the only viable commercial options for potential customers 20 of R&R platforms were Bazaarvoice and PowerReviews. GX683 at -339. Other 21 PowerReviews executives also recognized the duopolistic nature of the market when they 22 described their firm as “the #2 company in a two-horse race (with Bazaar[v]oice, out of 23 Austin, TX).” GX636 at -834. Consistent with its view that the R&R platform market was 24 a duopoly, PowerReviews noted that a merger of Bazaarvoice and PowerReviews would 25 eliminate competition by creating a “[m]onopoly in the market” when evaluating the 26 anticipated benefits of the acquisition. See, e.g., GX612 at 28. 27 28 44 iii. 1 2 100. The Industry Recognized R&R Platforms as a Separate Market Another factor supporting this market definition is that competitors and former 3 competitors in the R&R platform market recognized that R&R is a distinct market and that 4 after the merger there is now “one dominant player” in the R&R platforms space. GX75 5 (Moog Dep. 104:15-105:2, 105:5, 106:19-20, 106:23-107:3); see also GX952* at -258; 6 GX72* (Lithium Dep. 133:6-21); see Brown Shoe, 370 U.S. at 325 (“practical indicia,” 7 including industry or public recognition of the market, may determine boundaries of 8 relevant market). 9 101. Industry analysts recognize that R&R platforms comprise a distinct market. Gartner, a leading industry research analyst, noted that “[c]ustomer product reviews are the 11 United States District Court Northern District of California 10 most important tool that consumers use in determining where to shop. As such, customer 12 product reviews have become a must-have feature for multichannel retailer e-commerce 13 sites . . . . PowerReviews is one of the two dominant providers of community review 14 applications for multichannel and pure e-commerce retailers.” GX271 at -732, -737, -742; 15 see also GX604 at -404; GX1017 at -579-580. Gartner noted that through the merger 16 “Bazaarvoice . . . will acquire its sole rival, PowerReviews,” and further noted that 17 PowerReviews legacy clients should “[p]repare for potential migration to Bazaarvoice 18 products and potential price increases to cover the cost of this transaction as well as 19 historically higher-priced Bazaarvoice offerings.” GX1214 at -133-34. 20 102. Equity analysts also note the absence of other meaningful competitors in the R&R 21 platform market. For example, following Bazaarvoice’s IPO, Credit Suisse described 22 Bazaarvoice as “the market leader by at least a factor of two in an apparent duopoly for the 23 ratings and review space.” GX604 at -401; see also GX1017 at -579-80; GX687 at -075, - 24 091, -105 (Credit Suisse, Piper Jaffray, Deutsche Bank). 25 26 27 28 45 iv. 1 2 103. Other “Social Commerce” Products Are Not in the Market At trial, Bazaarvoice asserted that other social commerce products are competitive 3 with R&R and should be included in the product market. The evidence did not support 4 that assertion. 5 104. The user-generated-content collected by other social commerce products like forums, Q&A platforms, blogs, microblogs, and social networks, such as Facebook, serve 7 a different purpose than R&R and are not substitutes for R&R. Trial Tr. 461:5-462:7 8 (Cunningham/BJ’s Wholesale). None of those products provides potential consumers with 9 product-specific feedback from other consumers at the point of purchase. They are often 10 focused on brand advertising rather than driving the sale of individual products. GX169 11 United States District Court Northern District of California 6 (Orchard Supply Dep. 11:5-22, 22:6-10, 22:12-18, 22:20-24:3). In some cases, other 12 social commerce tools are used to manage customer service inquiries after a product 13 purchase. See, e.g., GX131 (Dillard’s Dep. 31:18-34:17). Customers on social media 14 sites, such as Facebook and Twitter, are “not in an eCommerce state of mind, so to speak, 15 [such customers are] not in the mind to purchase . . . .” GX131 (Dillard’s Dep. 33:3-23); 16 see also GX193 (Walgreens Dep. 73:24-74:11, 74:14-18, 74:20-22, 74:24-25, 75:2-5, 75:8, 17 75:10-11, 75:13, 75:15-21, 75:24); GX179 (REI Dep. 54:14-55:5, 55:8-15). Neither do the 18 other social commerce products collect the same type of structured, product-level data that 19 is associated with R&R. GX840 at -941. 20 105. Products that influence consumers immediately prior to making the purchase 21 decision, as R&R does, are fundamentally different than products that attempt to influence 22 consumer behavior at other times. See GX82 (Barton Dep. 81:7-82:20, 82:22-83:7); Trial 23 Tr. 284:18-285:23 (Levin/Clorox). Online retailers do not want consumers considering a 24 product to leave the retailer’s site in favor of a social site because the consumer may not 25 return to the retail site to finish the transaction. GX986* at 14; GX168 (Onlineshoes.com 26 Dep. 31:4-19); GX181 (Sears Dep. 60:5-19). 27 28 106. Because R&R platforms and other social commerce products serve different purposes, retailers and manufacturers often use R&R platforms in combination with one or 46 1 more other social commerce products. GX87 (Godfrey CID Dep. 33:1-12); GX90 2 (Luedtke Dep. 94:14-20); GX142 (Green Mountain Coffee Dep. 20:3-21:6, 22:7-23:17). 3 Many online retailers have R&R in addition to a Facebook page and Twitter account. 4 Most retailers with a Facebook page and a Twitter account believe that is competitively 5 necessary to also have R&R. ; GX143 (Guess? Dep. 49:8-11, 49:14-15); Trial Tr. 461:5-24 (Cunningham/BJ’s 6 7 Wholesale). In other words, other social commerce products are more complements to 8 R&R than substitutes. Pacitti agreed, stating that “[w]ithin the umbrella of social 9 commerce engagement tools there are also companies that are not directly competitive with one another on the basis of the features and functionality that they offer to their clients,” 11 United States District Court Northern District of California 10 including “Bazaarvoice’s ratings and review product.” GX492 (Pacitti Dep. 81:4-15, 12 82:10-15). 13 107. For example, Facebook is an online social network used by retailers and 14 manufacturers to communicate with customers and generate brand awareness. But it 15 serves a different purpose than R&R. Trial Tr. 461:516 (Cunningham/BJ’s Wholesale). 16 From a commercial standpoint, Facebook is more conducive to building brand awareness 17 than driving purchasing decisions. GX130 (Dick’s Sporting Goods Dep. 25:9-23, 26:1); 18 GX166 (OneCall Dep. 36:2-24); GX193 (Walgreens Dep. 70:1-71:12, 71:20-22, 72:2-9, 19 72:21-22, 72:25, 73:2-11); GX192 (Vitamin Shoppe Dep. 44:5-45:7, 45:9-14); GX77 20 (Heverley/BFG Dep. 31:5-8, 31:11-18); GX105 (Belk Dep. 25:24-26:22); GX179 (REI 21 Dep. 52:4-55:5, 55:8-56:3); GX110 (Blue Nile Dep. 38:1-39:18, 39:23-40:20); see also 22 GX986* at 13-14. Individual product reviews are generally found on a retailer’s website 23 rather than on its Facebook page. GX157 (K-Swiss Dep. 21:18-20, 21:22-22:3). Retailers 24 and manufacturers use Facebook as a complement to R&R; they do not substitute R&R 25 with a Facebook page. GX76 (Waltzinger/BBBeyond Dep. 38:23-39:7, 39:10-23); GX130 26 (Dick’s Sporting Goods Dep. 41:8-13, 41:16); GX193 (Walgreens Dep. 73:12-14, 73:16, 27 73:18-19, 73:22); GX98 (American Eagle Dep. 29:20-23, 30:1-8, 30:13); GX197 (XO 28 Group Dep. 51:25-52:7, 52:9-13, 52:16); GX160 (Lord & Taylor Dep. 36:9-17, 37:1-5, 47 1 37:8-9, 37:11-12, 37:16); GX170 (Overstock.com Dep. 66:2-4, 66:6); GX78 2 (Cunningham/BJ’s Wholesale Dep. 28:16-18, 29:4-7, 29:10, 29:12-21); GX179 (REI Dep. 3 58:15-59:2). Indeed, Bazaarvoice and PowerReviews viewed Facebook as a complement 4 to their R&R platforms and developed ways for their respective R&R platforms to 5 integrate with Facebook. GX82 (Barton Dep. 142:18-144:10); GX655 at -916; GX392 at - 6 104; GX530; GX88 (Hossain Dep. 96:4-5, 96:8-97:3). 7 108. In its pre-acquisition documents, Bazaarvoice states that companies that enhance 8 retailers’ and manufacturers’ use of Facebook, like Buddy Media, do not compete with 9 Bazaarvoice. See GX837 at -077-78; GX85 (Defossé Dep. 114:4-114:13, 114:15-114:20); Trial Tr. 891:20-892:3 (Collins). In a Form 10-Q filed with the SEC on March 14, 2013, 11 United States District Court Northern District of California 10 Bazaarvoice states that it does not yet face competition from Facebook and that Facebook 12 “do[es] not currently focus on our market.” GX969 at 36. At trial, Barton testified that he 13 was not aware of a situation where Bazaarvoice lost a customer to Facebook based on a 14 customer’s choice to use Facebook to the exclusion of Bazaarvoice. Trial Tr. 641:6-9 15 (Barton). Pacitti testified that Facebook does not compete “in any market in which 16 Bazaarvoice competes today. . . .” GX492 (Pacitti Dep. 44:8-13). Similarly, 17 PowerReviews did not consider Facebook a competitor. GX88 (Hossain Dep. 96:11-97:3). 18 109. Google+ is another social networking site. Google+ does not offer a competing 19 R&R platform. GX82 (Barton Dep. 170:1-3, 170:8-171:13). Google+ also does not 20 provide the same functionality as R&R and, like Facebook and other social commerce 21 products, online retailers and manufacturers would not abandon R&R in favor of Google+. 22 GX181 (Sears Dep. 62:13-63:11); Trial Tr. 1695:17-20, 1697:7-10 (Massuda/Sears); 23 GX179 (REI Dep. 57:4-5, 58:15-59:2). 24 110. Instagram and Pinterest are photo-sharing and social networking websites. 25 Instagram and Pinterest do not provide the same functionality as R&R and, like Facebook 26 and other social commerce products, online retailers would not abandon R&R in favor of 27 Instagram or Pinterest. GX98 (American Eagle Dep. 31:2-4, 31:7-13, 31:14-16, 31:18-21, 28 31:24); GX195 (Wine.com Dep. 34:25-35:5, 35:7-11); GX111 (Bon-Ton Dep. 45:4-14); 48 1 GX78 (Cunningham/BJ’s Wholesale Dep. 30:17-31:3, 31:7-12); GX179 (REI Dep. 57:21- 2 58:17). 3 111. YouTube is a video-sharing website. Online retailers would not consider 4 abandoning R&R functionality in favor of YouTube videos. GX181 (Sears Dep. 59:9-12, 5 64:15-19); Trial Tr. 1696:23-1697:2 (Massuda/Sears); GX98 (American Eagle Dep. 31:25- 6 32:3, 32:5-6, 32:20-22, 33:5). 7 112. Microblogs are not substitutes for R&R. Twitter is one of the most prominent microblogging sites. Twitter allows users to post “tweets” that followers of the poster can 9 read on their Twitter “feed.” GX986* at 11-12. Businesses often use Twitter to engage 10 with their customers and create brand awareness, but Twitter serves a different purpose 11 United States District Court Northern District of California 8 than R&R. GX130 (Dick’s Sporting Goods Dep. 26:5-9, 26:12-20); GX192 (Vitamin 12 Shoppe Dep. 45:15-18, 45:20-46:13, 46:15-17); GX77 (Heverley/BFG Dep. 31:5-8, 31:11- 13 18); GX147 (HSN Dep. 49:22-51:1, 51:4-52:21); GX195 (Wine.com Dep. 33:4-15, 33:18- 14 25, 55:10-11, 55:14-15); GX179 (REI Dep. 55:10-15, 56:22-57:3); GX153 (Jockey Dep. 15 37:24-38:4, 38:6-7, 38:9-11, 38:14-16). 16 113. Individual product reviews are generally found on a retailer’s website, not on its 17 Twitter feed. GX171 (Pacific Sunwear Dep. 37:25-38:1, 38:3-15, 38:17-39:9); GX109 18 (Blinds.com Dep. 84:16-86:1, 86:4). Retailers and manufacturers use Twitter as a 19 complement to R&R and would not substitute R&R with Twitter. GX181 (Sears Dep. 20 59:13-16, 64:2-6, 64:8-14); Trial Tr. 1696:12-22 (Massuda/Sears); GX130 (Dick’s 21 Sporting Goods Dep. 41:17-22, 41:25); GX141 (Golfsmith Dep. 51:17-18, 51:20-52:3, 22 52:5); GX193 (Walgreens Dep. 73:12-14, 73:16); GX98 (American Eagle Dep. 30:14-17, 23 30:20-23, 31:1); GX178 (Redbox Dep. 62:18-24, 63:1-4, 63:7); GX160 (Lord & Taylor 24 Dep. 37:17-38:1, 38:5-8); GX78 (Cunningham/BJ’s Wholesale Dep. 30:1-6, 30:10, 30:12- 25 16); GX179 (REI Dep. 58:15-59:2). PowerReviews did not consider Twitter a competitor. 26 GX88 (Hossain Dep. 97:4-19). 27 28 114. Blogs, online forums, and Q&As serve different functions than R&R. As used in eCommerce, blogs are a way to build brand awareness and interact with customers outside 49 1 of a commercial setting. Retailers and manufacturers use blogs as complements to R&R 2 and would not substitute R&R with a blog. GX183 (Shoplet Dep. 53:21-54:18); 3 (reviewed in camera); GX181 (Sears Dep. 4 59:17-20); Trial Tr. 1697:3-6 (Massuda/Sears); cf. GX170 (Overstock Dep. 66:12-67:7). 5 115. Retail blogs allow additional editorial content to be provided to customers and are 6 distinct from the UGC available through R&R. GX195 (Wine.com Dep. 34:3-18, 34:21- 7 24, 55:17-19); GX72* (Lithium Dep. 22:24-23:17, 32:9-15). Indeed, a company’s prior 8 adoption of a blog for its online retail site was viewed by Bazaarvoice as a “positive” in 9 trying to identify early customers. GX82 (Barton Dep. 32:6-23); Trial Tr. 627:14-17 10 United States District Court Northern District of California 11 (Barton). 116. Online forums allow users to interact and provide assistance in using a given 12 product in a different manner than the UGC provided in R&R. GX72* (Lithium Dep. 13 26:20-27:3, 28:9-19, 29:7-9, 29:12-30:22). Retailers and manufacturers do not view 14 forums as substitutes for R&R. GX77 (Heverley/BFG Dep. 46:21-24, 47:3-8); Trial Tr. 15 1623:12-14 (Ahmed/Amazon.com). 16 117. Q&A tools allow customers to ask questions that may not be answered by the 17 existing product description provided by the retailer or manufacturer on the product page. 18 Q&As complement the functionality and content provided by R&R. GX89 (Hurt Dep. 19 331:17-332:2); GX81 (Collins Dep. 134:8-13); GX109 (Blinds.com Dep. 46:2-8, 64:2-4, 20 64:6-25, 65:3-12); GX192 (Vitamin Shoppe Dep. 40:23-43:25, 44:4, 70:17-71:19, 71:23- 21 72:2, 72:4-9, 72:11); GX181 (Sears Dep. 57:13-58:18, 58:21-59:2); Trial Tr. 1694:8- 22 16(Massuda/Sears); GX146 (Home Depot Dep. 57:11-58:8); GX73* (Reevoo Dep. 10:13- 23 15, 10:23-11:20); GX72* (Lithium Dep. 23:21-24:20, 33:7-22); GX125 (Clorox Dep. 24 113:25-114:20); Trial Tr. 286:14-287:1 (Levin/Clorox); GX93 (Svatek Dep. 75:23-77:1, 25 77:10-78:13). But Q&A applications do not serve the same purpose as R&R platforms. 26 GX74* (TurnTo Dep. 116:8-11, 116:16-117:11); 27 28 . Rather, Q&As may fill the gaps left in customers’ minds after they read R&R. GX74* (TurnTo Dep. 107:3-6, 107:9-108:21). Bazaarvoice is not 50 1 aware of any customer that has replaced its R&R platform with a Q&A application. 2 GX1119 (Godfrey Dep. 24:2-6); see also Trial Tr. 1623:9-11 (Ahmed/Amazon.com). 3 Other Q&A providers do not consider Q&A a substitute for R&R, and they do not sell 4 their Q&A products as a substitute for R&R platforms. GX74* (TurnTo Dep. 21:14-16, 5 21:23-22:4, 22:8-24, 23:15-24, 24:3-13, 24:14-26:19, 26:22-24, 27:3-28:13, 28:16-19); 6 GX73* (Reevoo Dep. 9:2-5, 11:21-24, 12:6-20). 7 118. In short, other social commerce products do not compete directly with R&R 8 platforms, are not substitutes for R&R platforms, and do not constrain the prices of R&R 9 platforms. Trial Tr. 936:24-939:20 (Shapiro); GX983* at 48-51; GX74* (TurnTo Dep. 155:4-8, 155:11-157:13); GX492 (Pacitti Dep. 88:14-89:9). The Court disagrees with Dr. 11 United States District Court Northern District of California 10 Shehadeh’s view that the relevant product market is a cluster market of social commerce 12 tools providers given the distinctions in the tools described above. See, e.g., Brown Shoe, 13 370 U.S. at 325 (stating that a product market is “determined by the reasonable 14 interchangeability of use or the cross-elasticity of demand between the product itself and 15 substitutes for it”); Lucas Auto. Eng’g, Inc. v. Bridgestone/Firestone, Inc., 275 F.3d 762, 16 767 (9th Cir. 2001) (same). The tools may complement each other to some degree, but they 17 are sufficiently different, and R&R platforms are so critical for most customers that the 18 other tools are not adequate substitutes. Cf. F.T.C. v. Staples, Inc., 970 F. Supp. 1066, 19 1075 (D.D.C. 1997) (“the mere fact that a firm may be termed a competitor in the overall 20 marketplace does not necessarily require that it be included in the relevant product market 21 for antitrust purposes”). B. The United States Is the Relevant Geographic Market 22 23 119. The relevant geographic market is the location “where, within the area of 24 competitive overlap, the effect of the merger on competition will be direct and immediate.” 25 United States v. Philadelphia Nat’l Bank, 374 U.S. 321, 357 (1963). The “area of effective 26 competition in the known line of commerce must be charted by careful selection of the 27 market area in which the seller operates, and to which the purchaser can practicably turn 28 for supplies.” Id. at 359 (citation omitted). “The criteria to be used in determining the 51 1 appropriate geographic market are essentially similar to those used to determine the 2 relevant product market.” Brown Shoe, 370 U.S. at 336. 3 120. Bazaarvoice contended at trial that technology knows no borders and that the 4 geographic market for R&R should be worldwide. While R&R software is generally sold 5 worldwide “quite easily,” Trial Tr. 1906:20-22 (Shehadeh), there are several reasons why 6 the Court concludes that the United States is the relevant geographic market. 7 121. First, R&R requires more than code. R&R platform providers license their products for specific websites, often limited by geography. Manufacturers and retailers 9 who maintain separate websites for prospective consumers located in different countries 10 often have different managers and contracts for the different websites. See, e.g., Trial Tr. 11 United States District Court Northern District of California 8 1167:18-1168:7 (Friedland/Build.com); 12 (Caine Dep. 11:23-12:4, 63:14-17); GX92 (Osborne Dep. 39:8-16). Because R&R 13 platforms are provided under a software-as-a-service or cloud-based model, it is not 14 possible to resell the R&R platform to other websites, including the customer’s own 15 websites located in other territories. See, e.g., GX121* at -251, -258; GX437 at -049, - 16 054; GX369* at -611-12, -620; GX338* at -340; GX688* at -319. 17 122. ; GX79 Second, to offer a commercial syndication service in the United States, a R&R 18 platform provider must have a base of user-generated ratings and reviews written by 19 United States consumers. Few companies attempt to syndicate R&R between the United 20 States and other countries due to even slight differences in the products or their product 21 SKUs (stock-keeping units) across geographies that make it difficult to match ratings and 22 reviews with products across regions. GX87 (Godfrey CID Dep. 84:9-20); ; GX96* (Acer Dep. 73:4-14). 23 24 123. Third, understanding the nuances of language and culture matters in moderation. 25 American customers often prefer to have United States-based, native English speakers 26 perform the moderation of their R&R. GX248 at -595; GX90 (Luedtke Dep. 279:10-20, 27 280:9-14, 280:17-281:2, 281:13-14); Trial Tr. 371:8-372:2 (Luedtke); GX89 (Hurt Dep. 28 52 1 264:11-17, 264:20-24); Trial Tr. 106:7-17 (Hurt); cf. GX87 (Godfrey CID Dep. 84:21- 2 85:3). 3 124. Bazaarvoice’s conduct was consistent with the fact that there are separate 4 geographic markets. When expanding internationally, Bazaarvoice prioritized hiring a 5 local salesperson in that new country. GX82 (Barton Dep. 87:25-88:9); Trial Tr. 630:19- 6 632:6 (Barton); GX92 (Osborne Dep. 26:9-12, 27:17-28:7); GX81 (Collins Dep. 198:7- 7 23). Bazaarvoice’s United States sales personnel focus on selling to companies’ domestic- 8 facing websites, even while its sales personnel in other countries negotiate the sale of 9 Bazaarvoice’s R&R platform for the same companies’ foreign websites. GX92 (Osborne Dep. 38:9-21, 39:3-16); Trial Tr. 704:19-25 (Osborne). When announcing the 11 United States District Court Northern District of California 10 PowerReviews acquisition, Hurt said: “we found in serving large Fortune 500 clients that 12 it’s very important to serve them from a local presence. People want to deal with each 13 other from that local presence . . . .” GX840 at -943. Osborne testified that Bazaarvoice 14 faced different competitive dynamics in different geographies. Trial Tr. 705:5-7 15 (Osborne). 16 17 125. The participants in the R&R market understand that there are separate geographic markets as well. . GX952* at -258. Hurt testified that prior to the 18 19 merger he believed that PowerReviews was Bazaarvoice’s biggest competitor “for the U.S. 20 market and ratings and reviews.” Trial Tr. 130:7-12 (Hurt); see also Trial Tr. 131:1, 21 131:19-20, 193:5-7, 231:14-17 (Hurt); cf. Trial Tr. 665:17-666:5, 682:23-25 (Barton) 22 (discussing Reevoo’s history as a “European competitor”). Retailers and manufacturers 23 concur that there are separate markets. See, e.g., GX146 (Home Depot Dep. 67:3-68:4); 24 GX125 (Clorox Dep. 121:25-123:4). 25 C. Economic Analysis Confirms that R&R Platforms Sold to United States Customers Is the Relevant Market 26 27 28 126. The government’s expert witness, Dr. Carl Shapiro, evaluated the likely competitive effects of the acquisition of PowerReviews by Bazaarvoice. Dr. Shapiro 53 1 earned his Ph.D. in economics from M.I.T. and specializes in the field of industrial 2 organization generally and antitrust specifically. He has published extensively. He twice 3 served as the Deputy Assistant Attorney General for Economics in the Antitrust Division 4 of the United States Department of Justice. He was one of three members of the 5 President’s Council of Economic Advisors. Dr. Shapiro is currently the Transamerica 6 Professor of Business Strategy at the Haas School of Business at the University of 7 California, Berkeley. He has considerable experience in the application of economics for 8 the purpose of enforcing antitrust laws and is co-author of a book with Hal R. Varian, 9 Information Rules: A Strategic Guide to the Network Economy, which examines 10 United States District Court Northern District of California 11 competitive strategy in high-tech industries. 127. Dr. Shapiro applied the MERGER GUIDELINES, jointly promulgated by the Antitrust 12 Division of the United States Department of Justice and the Federal Trade Commission, in 13 order to assess the likely effects of the merger. It suggests identifying relevant markets, 14 consisting of a relevant product market and relevant geographic market, to determine the 15 areas where a substantial lessening of competition may be observed. While determining 16 whether a substantial lessening of competition may occur “need not start with market 17 definition,” “market definition can be informative regarding competitive effects.” MERGER 18 GUIDELINES § 4. In this regard, Dr. Shapiro testified that the economic framework for 19 analyzing mergers is the same for consummated and unconsummated mergers. Trial Tr. 20 900:13-21 (Shapiro). He opined that the relevant product market is R&R platforms for 21 retailers and manufacturers and the relevant geographic market is the United States. Trial 22 Tr. 901:10-19, 936:17-23 (Shapiro). 23 128. Courts often employ a “hypothetical monopolist” test as an analytical method to aid 24 the determination of the relevant product market. See, e.g., H & R Block, Inc., 833 F. 25 Supp. 2d at 51-52; Oracle Corp., 331 F. Supp. 2d at 1112. The hypothetical monopolist 26 test is endorsed by the MERGER GUIDELINES. Dr. Shapiro applied the hypothetical 27 monopolist test to define the relevant product market. See, e.g., GX984* at 1. 28 54 1 129. The purpose of the hypothetical monopolist test is to assess whether there would be 2 significant customer harm if one firm controlled all of the products in the candidate market 3 relative to pre-merger competition. Trial Tr. 901:20-902:11 (Shapiro). The test asks 4 whether a profit-maximizing hypothetical monopolist would impose at least a small but 5 significant non-transitory increase in price of five to 10 percent (“SSNIP”) on a proposed 6 candidate set of products. GX983* at 40-41; MERGER GUIDELINES § 4.1; see also Trial Tr. 7 901:20-903:16 (Shapiro). 8 130. In this case, the question is whether enough current users of R&R platforms would drop R&R platforms in response to a SSNIP to make a five to ten percent price increase 10 unprofitable. Trial Tr. 901:20-903:16 (Shapiro); GX983* at 42. Dr. Shapiro noted, “the 11 United States District Court Northern District of California 9 smaller the number of customers who would shift from products in the candidate market to 12 products outside that market in response to a SSNIP imposed on products in the candidate 13 market, the more likely that market will satisfy the test.” GX983* at 42. The hypothetical 14 monopolist will impose a SSNIP on a product if the profit gained from the SSNIP 15 outweighs the total profit lost from customers who will no longer purchase the product at 16 the higher price and will not switch to another product in the market. GX983* at 45. 17 131. The MERGER GUIDELINES also provide for defining the relevant geographic market 18 based on the customer location “[w]hen the hypothetical monopolist could discriminate 19 based on customer location.” MERGER GUIDELINES § 4.2.2; cf. Malaney v. UAL Corp., 10- 20 cv-02858-RS, 2010 WL 3790296, at *6 (N.D. Cal. Sept. 27, 2010), aff’d, 434 F. App’x 21 620 (9th Cir. 2011) (“In other words, rather than a snapshot view of customer choices in a 22 particular community, the geographic market is the area where customers can turn in 23 response to a ‘hypothetical monopolist’ increasing prices.”). 24 132. As described by Dr. Shapiro, in the R&R platform market, a hypothetical 25 monopolist can identify and target price increases to a particular group of customers based 26 on geography because those customers cannot engage in arbitrage with customers in 27 another geographic area. The relevant geographic market is defined around those 28 customers to whom a hypothetical monopolist would profitably impose a SSNIP. GX983* 55 1 at 41-42. A supplier in a different area “that has provided or attempted to provide” a 2 relevant product to customers in the geographic market should be included in the market. 3 GX983* at 42; GX984* at 4. In applying the hypothetical monopolist test, Dr. Shapiro 4 asked, “[i]f a single firm controlled all of the ratings-and-reviews software to customers in 5 the United States, would that be a significant harm to customers?” Trial Tr. 924:22-24 6 (Shapiro). Here, the hypothetical monopolist would control Bazaarvoice, PowerReviews, 7 all in-house solutions, and the fringe suppliers, and Dr. Shapiro concluded that this would 8 cause a significant harm to customers. Trial Tr. 924:25-925:6 (Shapiro). 9 133. Dr. Shapiro identified several factors to support R&R platforms as a relevant product market. First, it is unlikely that customers would substitute other products for 11 United States District Court Northern District of California 10 R&R platforms. Having a R&R platform is “table stakes” that has become “a necessary 12 part of being a major online retailer.” Trial Tr. 926:3-5 (Shapiro). The trend among top 13 retailers to use R&R suggests that R&R platforms have become a necessary component of 14 an online retail website. See Trial Tr. 926:25-927:17 (Shapiro). Bazaarvoice’s economic 15 expert, Dr. Shehadeh, agreed that many customers are “unlikely to consider giving up” 16 R&R. Trial Tr. 1721:14-15 (Shehadeh). 17 134. That many customers testified that they would not abandon R&R platforms (or 18 even their current R&R platform provider) in the face of a price increase is also consistent 19 with Dr. Shapiro’s conclusion that R&R platforms are a relevant product market. See Trial 20 Tr. 1195:17-20 (Friedland/Build.com) (Build.com would not drop R&R if its price 21 increased by five percent); GX983* at 43-44; GX130 (Dick’s Sporting Goods Dep. 43:11- 22 16); GX193 (Walgreens Dep. 83:22-23, 84:1-2, 84:4-5, 84:8-12, 84:14-17, 84:20-21, 85:5- 23 14, 85:20-21); GX145 (h.h. gregg Dep. 23:23-24:7); GX111 (Bon-Ton Dep. 37:18-38:7); 24 GX76 (Waltzinger/BBBeyond Dep. 32:20-21, 32:25-33:9, 34:22-35:2, 37:14-16, 37:20- 25 38:10); Trial Tr. 566:4-11, 567:5-7, 567:25-568:12 (Waltzinger/BBBeyond); GX109 26 (Blinds.com Dep. 66:7-11, 66:13-16); GX131 (Dillard’s Dep. 72:23-73:18); see also 27 GX1106* at -617 (customer agreed to 28 increase in service or functionality). Dr. Shapiro concluded the evidence was clear that percent price increase post-merger without an 56 1 “no or almost none of them would” drop their R&R platforms if faced with such a price 2 increase. Trial Tr. 926:11-13 (Shapiro). 3 135. Second, the companies’ pre-merger ROI analysis supports this product market. 4 Trial Tr. 926:14-24 (Shapiro). As part of their sales process, Bazaarvoice and 5 PowerReviews typically prepared ROI analyses for potential customers to convince them 6 to adopt their R&R platform. GX983* at 43-44. Bazaarvoice’s ROI estimates of the 7 substantial benefits of R&R to manufacturers and retailers compared to the cost of the 8 product are further evidence that a small but significant nontransitory increase in price 9 “will not cause customers to drop ratings and reviews.” Trial Tr. 926:22 (Shapiro). 10 136. Third, the companies’ deal documents support the product market. Trial Tr. United States District Court Northern District of California 11 929:11-17 (Shapiro). The documents indicate that both companies’ executives believed 12 the merger would avoid price erosion and allow the merged firm to profitably raise prices 13 at least five percent above pre-merger levels. Trial Tr. 930:2-8 (Shapiro). Bazaarvoice 14 and PowerReviews each asserted that the merged firm could increase prices post-merger. 15 Someone with 100 percent of the market could do so even more profitably. Trial Tr. 16 930:9-12 (Shapiro). Indeed, this was underscored by Christian Friedland, the founder of 17 Build.com, who admitted that if Bazaarvoice tried to raise its price by five percent, it 18 “might be an empty threat” to threaten to cancel the contract. Trial Tr. 1188:12-1189:12 19 (Friedland). 20 137. Fourth, in head-to-head competition, Bazaarvoice and PowerReviews often 21 discounted their R&R platform prices in excess of the five to ten percent applied by the 22 SSNIP test. The hypothetical monopolist would not face the same pressure to discount its 23 prices and the customer would still purchase R&R platforms, in essence accepting a 24 SSNIP. GX983* at 42-43. 25 138. Finally, Dr. Shapiro looked at the recapture rate to support his opinion. In order for 26 a SSNIP imposed by a hypothetical monopolist to be profitable, the recapture rate (“the 27 percentage of sales lost by one product in the candidate market, when its price alone rises, 28 that is recaptured by other products in the candidate market”) must be high enough that the 57 1 incremental profits from the increased price plus the incremental profits from the 2 recaptured sales that goes to other products in the candidate market exceeds the profits lost 3 to products outside the candidate market. GX983* at 45; MERGER GUIDELINES § 4.1.3. He 4 concluded that the minimum recapture rate needed for R&R platforms to form a relevant 5 market would be 17 percent. Trial Tr. 933:5-19 (Shapiro); GX983* at 46. Dr. Shapiro 6 estimated various possible recapture rates and proxies for the recapture rate and determined 7 that all variations were substantially larger than 17 percent. Trial Tr. 935:13-936:14 8 (Shapiro); GX983* at 45-48. He performed another estimation of the recapture rate based 9 upon data cited by Bazaarvoice’s expert and also found that the recapture rate is well above the 17 percent threshold necessary to show R&R platforms to be a relevant market. 11 United States District Court Northern District of California 10 GX984* at 3. 12 139. Dr. Shapiro also applied the hypothetical monopolist test to determine the relevant 13 geographic market. Where, as here, a hypothetical monopolist could price discriminate, 14 i.e., set different prices for different customers based on customer location, the geographic 15 market is based on the location of the customers, not the suppliers. MERGER GUIDELINES § 16 4.2.2. Dr. Shapiro concluded that to properly assess the competitive effects of the merger, 17 the relevant market should be defined as the United States, specifically domestic retailers 18 and manufacturers. Trial Tr. 943:7-11 (Shapiro). 19 140. Dr. Ramsey Shehadeh testified as Bazaarvoice’s economic expert. He analyzed the 20 merger and evaluated its likely competitive effects. Dr. Shehadeh earned his Ph.D in 21 economics from Cornell University and works with NERA Economic Consulting, where 22 he is Senior Vice President, Partner, and Chair of the Global Competition Policy Practice. 23 He has testified about and written widely on economic topics relevant to this case, and is 24 Vice Chair of the Economics Committee of the Section of Antitrust Law of the American 25 Bar Association. 26 141. Dr. Shehadeh argued that the government offered no basis to limit the relevant 27 product market to R&R only for manufacturer and retailer customers. Other customers 28 that use R&R platforms fall outside this market, such as airlines (e.g., Southwest Airlines), 58 1 travel companies (e.g., International Hotels Group, Starwood, Expedia, and Princess 2 Cruises), and financial services firms (e.g., First Tennessee Bank). However, customers 3 from these other sectors contributed little to the revenue base of Bazaarvoice or 4 PowerReviews. All of Bazaarvoice’s customers use R&R and at least 88 percent of its 5 revenue was from retailers and manufacturers. Trial Tr. 273:20-274:10 (Hurt). Around 90 6 percent of PowerReviews’s revenue came from retailers. Trial Tr. 442:6-12 (Luedtke). 7 Dr. Shapiro also compared use of R&R in the Fortune 500 companies, including 8 companies that were not retailers and manufacturers, and demonstrated the same basic 9 market structure. Accordingly, any problem in limiting the market to manufacturers and 10 United States District Court Northern District of California 11 retailers is de minimis. 142. Bazaarvoice contends that Dr. Shapiro’s product market analysis is flawed because 12 Dr. Shapiro defines the relevant product market by reference to demand considerations 13 alone and does not include supply elasticity in his market definition analysis. Trial Tr. 14 1722:4-16 (Shehadeh); GX0984 at 2-3; DX1736 at ¶ 128. In support, Bazaarvoice cites 15 Rebel Oil Co., Inc. v. Atlantic Richfield Co., 51 F.3d 1421, 1436 (9th Cir. 1995), for the 16 proposition that “[w]hile economic theory might not demand that supply substitution be 17 considered in defining the market, Ninth Circuit precedent does.” Bazaarvoice Post-Trial 18 Brief at 25. 19 143. A rapid entrant is a company that is not currently selling a R&R platform, but could 20 enter the market very quickly and easily without making a substantial investment or 21 incurring risk. Trial Tr. 970:1-13, 972:7-10 (Shapiro); MERGER GUIDELINES § 5.1. Dr. 22 Shapiro testified that considering rapid entrants at this stage of the analysis will lead to 23 “errors.” Trial Tr. 940:23-941:5 (Shapiro). Product market definition focuses on the 24 demand side and asks “what do the customers need.” Trial Tr. 941:4 (Shapiro). The 25 “standard” merger analysis, which Dr. Shapiro employed, considers supply-side 26 substitution, including rapid entrants, when calculating market shares and analyzing entry. 27 Trial Tr. 940:9-941:5 (Shapiro); MERGER GUIDELINES § 4. Dr. Shehadeh undercut 28 Bazaarvoice’s argument when he agreed that it is acceptable to consider supply side 59 1 substitution when considering market share as opposed to market definition. Trial Tr. 2 1722:4-12 (Shehadeh) (“There’s nothing wrong with that approach in theory, considering 3 it at that second stage.”). The Court disagrees with an interpretation of Rebel Oil which 4 would require consideration of rapid entrants when defining the market, for the reasons 5 described later in the Discussion section of the Opinion. Given Dr. Shapiro’s (and the 6 Court’s) conclusion that there are no rapid entrants anyway, Bazaarvoice’s criticism is 7 immaterial. 8 9 144. Bazaarvoice also contends that Dr. Shapiro’s product market analysis is unreliable because (1) no data exist to perform the hypothetical monopolist test for R&R, Trial Tr. 1733:12-1735:24 (Shehadeh); (2) his recapture analysis does not establish that a 11 United States District Court Northern District of California 10 hypothetical monopolist could impose a SSNIP on R&R solutions because it makes no 12 attempt to model how customers would behave in response to a price increase and only 13 examines actual switching behavior absent a price incentive to switch due to an increase in 14 R&R prices relative to all other products, DX1736 at ¶ 183; (3) he does not calculate a 15 recapture rate for the market, but instead offers a “proxy” based on what IR500 customers 16 did at prevailing prices, Trial Tr. 935:13-936:15 (Shapiro); (4) Bazaarvoice and 17 PowerReviews do not track revenue for R&R separately from revenue for their other 18 products, so the margin used in Dr. Shapiro’s recapture analysis reflects all social 19 commerce products sold by Bazaarvoice and PowerReviews, Trial Tr. 1083:25-1084:14 20 (Shapiro); GX80 (Bolian Dep. 49:22-50:7); DX1802 at 65; (5) his analysis of a 21 hypothetical price increase on PowerReviews’s Enterprise product is meaningless because 22 it incorrectly assumes that differences in annual service fees (“ASF”) between customers 23 can be attributed to R&R solutions specifically and therefore relies on price comparisons 24 that lack any foundation, DX1736 at ¶ 184; and (6) his recapture analysis shows it would 25 be unprofitable for Bazaarvoice to raise prices post-merger because the number of IR500 26 customers leaving Bazaarvoice for solutions other than PowerReviews would discipline 27 any price increase. Trial Tr. 1881:3-19, 1888:15-23 (Shehadeh). 28 60 1 145. Some of Bazaarvoice’s criticisms ring true. The data that exist regarding this 2 market are imperfect, in part because the market is relatively new. But that does not mean 3 that one should ignore the existing data or the market realities. For example, there is no 4 real dispute that the overwhelming majority of PowerReviews’s and Bazaarvoice’s 5 business came from R&R, so to complain that Dr. Shapiro did not distinguish between 6 revenue from R&R and other the parties’ other offerings that complement R&R is 7 unconvincing. Dr. Shehadeh did not try to perform the hypothetical monopolist test in 8 expanding the relevant product market to include these other social commerce tools, Trial 9 Tr. 937:5-15 (Shapiro). While the data available to Dr. Shapiro may not have been perfect, it sufficiently reflected the state of the market as shown by other evidence in this case. 11 United States District Court Northern District of California 10 And while other social commerce tools provide sites with user generated content and may 12 help increase conversions and improve traffic, witness after witness testified that R&R was 13 nonetheless essential to drive conversion. The other social commerce tools are most often 14 complements rather than substitutes, and there is no persuasive evidence that this will 15 change in the next two years. 16 146. Bazaarvoice objected that although Dr. Shapiro concluded the geographic market is 17 the United States, he included in the market foreign suppliers that serve domestic 18 customers. Trial Tr. 941:6-25 (Shapiro). In counting foreign suppliers, Dr. Shapiro only 19 considered their success in the United States in order to meaningfully measure their 20 significance to customers with domestic-facing websites. Trial Tr. 942:1-8 (Shapiro). Dr. 21 Shapiro explained that because foreign suppliers’ R&R platforms cannot be resold to 22 domestic customers, it would make no sense to include foreign suppliers who are not 23 selling into the United States. Trial Tr. 942:17-943:6 (Shapiro). They do not affect the 24 prices charged to domestic customers. 25 147. Dr. Shapiro’s analysis confirms what was apparent from the non-expert testimony: 26 other social commerce tools, including social networking sites, Q&As, and forums, either 27 serve a different purpose than R&R or are insufficient substitutes such that customers 28 would not switch from R&R to a social commerce tool in the face of a SSNIP. GX983* at 61 1 48-51. Dr. Shapiro’s testimony buttressed the findings of the Court that R&R platforms in 2 the United States for retailers and manufacturers are the relevant product and geographic 3 markets. 4 5 V. ANALYSIS OF GOVERNMENT’S PRIMA FACIE CASE AND BAZAARVOICE’S RESPONSE 6 A. Bazaarvoice’s market share establishes a prima facie violation of Section 7 7 8 9 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 148. A firm’s market share is usually a primary measure of that firm’s presumed competitive significance. GX983* at 51; see Philadelphia Nat’l Bank, 374 U.S. at 321 (“Without attempting to specify the smallest market share which would still be considered to threaten undue concentration, we are clear that 30% presents that threat.”). Dr. Shapiro analyzed the market shares using several different metrics, each of which confirmed the commanding position occupied by Bazaarvoice and PowerReviews. Trial Tr. 943:25944:5 (Shapiro). 149. First, Dr. Shapiro used the IR500 to calculate market shares. Trial Tr. 944:6-11 (Shapiro). The IR500 provides the best available list of the largest internet retailers. Generally, online retailers with at least $15 million in revenues disclose their R&R platform provider to Internet Retailer, which creates the annual list. The IR500 accounts for approximately 80 percent of online retail commerce in the United States and Canada. Trial Tr. 1085:6-11 (Shapiro); DX1736 at 17. Although Bazaarvoice questioned the reliability of IR500 data at trial, the list and data are well-recognized by companies involved in eCommerce. Before the merger, Bazaarvoice and PowerReviews, in the regular course of business, closely tracked their IR500 market share. No other metric was as prominently mentioned in the pre-acquisition documents introduced at trial. 150. Bazaarvoice frequently cited its number of IR500 customers when promoting itself to potential customers and investors. GX1168 at -443; GX484 at -022. For example, in connection with its IPO, Bazaarvoice touted its share of IR500 customers in prepared responses to questions that it would likely face from potential investors: 62 [Question:] What type of market share do you currently have and what do you believe is a realistic share that you can get from a longterm perspective? 1 2 [Answer:] There are a couple of different ways that we look at market share and market penetration. First, we currently represent 145 of the top 500 Internet Retailers, which is close to 30% market share in US retail online but remember that a lot of our retail clients are multichannel and not all are listed in the IR 500. 3 4 5 6 7 8 9 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 GX425 at -919. Hurt testified that the list was one that “we would use in terms of measuring ourselves.” Trial Tr. 183:9-12 (Hurt). It also used the list to classify PowerReviews customers that it was targeting. GX39 at -934, -939; Trial Tr. 511:16513:9 (Defossé); see also Trial Tr. 1429:11-15 (Meredith) (referring to the IR500 in response to a question from the Court). 151. In its due diligence analysis of the PowerReviews acquisition, Bazaarvoice stressed that “[t]he enhanced consumer reach may be illustrated by the fact that a combined entity would provide solutions to 243, or 49%, of the 500 largest North American retailers by online sales listed on the 2012 Internet Retailer 500 (‘IR 500’) and immediately increase the IR 500 penetration of Bazaarvoice by 49%.” GX925 at -946. In an email to other Bazaarvoice executives in October, 2011, Collins relied on the IR500 as an indication of Bazaarvoice’s market share: According to the Internet Retailer, PR has 78 IR 500 clients with combined retail sales of $17BB. We have 136 and $54BB in retail sales. The combination would produce a truly massive audience for us and would definitively tip the scales in our permanent favor on the network front . . . . Our share of the IR 500 would move close to 50% in absolute terms and higher in retail sales coverage. 19 20 21 22 Bazaarvoice also pointed to the IR500 when touting the benefits of the acquisition. GX522 at -035. 23 24 152. Similarly, in the regular course of business prior to the merger, PowerReviews 25 regularly tracked the IR500 to see what percentage had implemented the PowerReviews 26 R&R platform. GX90 (Luedtke Dep. 33:5-9); GX632. Bazaarvoice executives testified 27 that the company considered the impact that the acquisition of PowerReviews would have 28 on Bazaarvoice’s IR500 market share. GX770 at -218; Trial Tr. 145:11-15 (Hurt); GX82 63 1 (Barton Dep. 103:7-17); Trial Tr. 635:13-636:3 (Barton); GX92 (Osborne Dep. 282:21- 2 283:6); GX88 (Hossain Dep. 36:5-37:9); GX492 (Pacitti Dep. 149:7-23, 150:1). 3 153. Customers and industry analysts also look to the information maintained by the 4 IR500 to determine who are the R&R platform vendors in the marketplace and as a useful 5 metric of market share and market leadership. GX130 (Dick’s Sporting Goods Dep. 6 30:17-32:7, 47:1-13, 47:22-48:1); GX108* (Black & Decker Dep. 81:1-20, 82:22-83:14); 7 GX78 (Cunningham/BJ’s Wholesale Dep. 19:24-20:16); GX1017 at -414, -584-85. 8 Although the IR500 only tracks retailers, the data are useful to indicate the competitive 9 significance of R&R vendors for brands because, when syndicating, brands care about how much commerce is taking place on the retailer’s site. Brands will be more likely to select 11 United States District Court Northern District of California 10 the R&R provider whose syndication network provides them access to the largest number 12 of retail consumers. Trial Tr. 964:7-15, 966:23-967:12 (Shapiro). Because of this, 13 retailers are very important to Bazaarvoice’s brand business. Bazaarvoice’s strategy was 14 to build a large base of retailers in order to attract brand customers. Trial Tr. 967:13-24 15 (Shapiro). 16 154. Based on IR500 customer count in 2012, Bazaarvoice had a market share of 40 17 percent of companies using R&R, PowerReviews had a market share of 28 percent, in- 18 house solutions comprised 28 percent of the market, and all other commercial vendors had 19 a combined market share of three percent. Trial Tr. 947:3-22 (Shapiro). Based on the 20 IR500 customer count in 2012, the combined Bazaarvoice-PowerReviews share is 68 21 percent of the companies using R&R. GX985* at Ex.14A; GX1062. 22 155. The market shares shown by the IR500 customer count indicate that there are a 23 number of low-quality alternatives. To assess their commercial success, the standard 24 methodology is to measure their market shares. Trial Tr. 950:3-7 (Shapiro). Their small 25 market share of three percent (15 customers) indicates that these firms are “unlikely to be 26 as competitively significant as firms with much bigger market share.” Trial Tr. 948:8-18 27 (Shapiro). “These [fringe] firms are not very attractive to most customers [because] 28 [t]hey’re not winning very much business.” Trial Tr. 949:15-24 (Shapiro). 64 156. 1 Dr. Shapiro also used the IR500 data to measure market shares based on “customer 2 revenues,” which reflects the relative volume of online commerce that is handled by the 3 various R&R platforms. Trial Tr. 963:8-20 (Shapiro). Customer revenue is one 4 appropriate way to measure market shares for the customers in the relevant market. Trial 5 Tr. 967:25-968:14 (Shapiro). Retailers and brands account for approximately 90 percent 6 of Bazaarvoice’s and PowerReviews’s revenues. Trial Tr. 968:15-22 (Shapiro); Trial Tr. 7 273:20-274:10 (Hurt); Trial Tr. 442:6-12 (Luedtke). Accordingly, customer revenue is a 8 useful metric to measure the significance of R&R platform competitors to both retailers 9 and brands. See Trial Tr. 964:4-15 (Shapiro). 157. 10 Based on customer revenues, Bazaarvoice has a market share of almost 41 percent United States District Court Northern District of California 11 of companies using R&R, PowerReviews has a market share of 15 percent, and all other 12 commercial vendors have a combined market share of two percent. In-house solutions 13 have a combined market share of 42 percent, but one customer—Amazon—accounts for 14 approximately 67 percent of that, or a 28 percent share of the market. The combined 15 Bazaarvoice-PowerReviews share is 56 percent. See Trial Tr. 963:8-964:1 (Shapiro); 16 GX985* at Ex. 15A; GX1063. 10 158. 17 Dr. Shapiro recognized that the IR500 data does not provide a “perfect measure” of 18 the R&R market, so he performed additional analyses to check the robustness of these 19 results. Trial Tr. 974:12-20 (Shapiro). First, he measured shares in the broader IR1000. 20 The additional 500 firms are small retailers with sales ranging between about $15 million 21 to several hundred thousand dollars. Trial Tr. 969:1-7 (Shapiro). (The IR500 reflects 98 22 percent of the revenue shown on the IR1000.) Dr. Shapiro found that no R&R provider, 23 other than the ones previously identified as providers for the IR500, served more than ten 24 of these extra 500 retailers. Trial Tr. 976:16-977:1 (Shapiro). 25 26 27 28 10 Dr. Shapiro’s initial market share calculations based on customer revenues contained numerous mistakes which required correction. These adjustments resulted in about a one percent adjustment to PowerReviews’ market share and about a two percent adjustment to Bazaarvoice’s market share. Trial Tr. 965:6-23 (Shapiro); compare GX1077 with GX1063. 65 1 159. Second, he measured shares using the list of companies in the Fortune 500. The 2 Fortune 500 is informative because it is a standard set of large companies that includes, but 3 is certainly not limited to, retailers and brands. Trial Tr. 975:6-18 (Shapiro). Based on the 4 Fortune 500, for those companies using R&R, Bazaarvoice has a market share of 67 5 percent, PowerReviews has a market share of 16 percent, in-house has a market share of 6 14 percent, and all other commercial vendors have a combined market share of about three 7 percent. Trial Tr. 976:1-15 (Shapiro); GX1078. 8 160. Third, he analyzed a sample that Dr. Shehadeh had created of 1,000 actual and prospective customer accounts that was randomly drawn from a pool of 29,494 workable 10 accounts worldwide in the Bazaarvoice Salesforce.com database. Trial Tr. 979:19-980:5 11 United States District Court Northern District of California 9 (Shapiro). After limiting the data set to domestic retailers and brands that have a R&R 12 platform, Dr. Shapiro found that in-house had a market share of 40 percent, Bazaarvoice 13 had a market share of 38 percent, PowerReviews had a market share of 21 percent, or a 14 combined total of 59 percent, and all others had a share of less than 2 percent. Trial Tr. 15 980:9-981:11 (Shapiro); GX1057. 16 161. Dr. Shehadeh reached a different conclusion concerning Bazaarvoice’s market 17 share. After he drew his random sample, he determined whether PowerReviews or 18 Bazaarvoice provided any solutions to that website. 87 accounts with no corresponding 19 webpage were excluded from the analysis. He concluded that the combined firm provides 20 service to only 11 percent of this sample. DX1433 at ¶¶ 252-253. Then Dr. Shehadeh 21 took into account that 51 percent of Bazaarvoice’s prospective customers implement some 22 social marketing solution already. He assumed that only those customers who have 23 implemented some social marketing solution would be candidates to purchase R&R. 24 Using this metric, Dr. Shehadeh concluded that Bazaarvoice and PowerReviews combined 25 provided solutions to 21 percent of the sample. DX1433 at Exhibits 8, 9. Dr. Shehadeh’s 26 conclusion was not credible because he did not further refine his analysis to include only 27 American companies that used R&R on their websites, and because he included, without 28 66 adequate justification, companies that had not yet implemented R&R. 11 1 162. 2 Dr. Shapiro’s final robustness check was to compile data on the total number of 3 customers for each of the commercial R&R vendors identified in Dr. Shehadeh’s report. 4 Trial Tr. 978:4-9 (Shapiro). The customer count information generally comes from 5 company depositions as well as Bazaarvoice’s due diligence memorandum regarding the 6 merger. GX1064*; Trial Tr. 978:4-17 (Shapiro). The results of Dr. Shapiro’s analysis are 7 presented in the chart below: 8 9 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 163. The two columns on the right reflect the vendor’s customer count and sales in the 20 IR500. The middle column, titled “PRR Customers,” reflects the vendor’s total customer 21 count, including customers in and outside the IR500. Trial Tr. 978:4-17 (Shapiro). The 22 results are similar to what was found in the other robustness checks. Companies other than 23 24 11 25 26 27 28 It is worth noting that, based on Dr. Shapiro’s raw count of which vendors provide R&R to IR500 companies—underlying figures which Bazaarvoice does not dispute—Bazaarvoice and PowerReviews combined provide R&R to 57 per cent of all IR500 companies, including those that do not have R&R. All other providers combined provide R&R to fewer than 4 per cent of IR500 companies. Approximately 25 per cent of IR500 companies have an in-house R&R solution. GX985*. Accordingly, even when including companies that do not provide R&R, the basic market structure is the same: Bazaarvoice and PowerReviews are the two dominant providers of R&R and no other provider comes close to their share of the market. 67 1 Bazaarvoice and PowerReviews have very few customers. 12 These market shares are 2 meaningful because R&R platform vendors with little or no market share are unlikely to 3 provide as strong a competitive check on Bazaarvoice as PowerReviews did before the 4 merger. GX984* at 7. 164. 5 Each of Dr. Shapiro’s checks revealed the same basic market structure: 6 Bazaarvoice and PowerReviews are the two largest companies; a significant number of 7 customers choose in-house solutions; and there are some small fringe competitors with 8 very little market share. Trial Tr. 982:11-19 (Shapiro). 165. 9 Bazaarvoice’s acquisition of PowerReviews significantly increased concentration in the already highly concentrated R&R platform market. The MERGER GUIDELINES place 11 United States District Court Northern District of California 10 great weight on increased market concentration in assessing the likely competitive effects 12 of mergers. In assessing market concentration, the MERGER GUIDELINES use the 13 Herfindahl-Hirschman Index (“HHI”). MERGER GUIDELINES § 5.3. See also F.T.C. v. 14 H.J. Heinz Co., 246 F.3d 708, 716 (D.C. Cir. 2001) (“Market concentration, or the lack 15 thereof, is often measured by the Herfindahl–Hirschmann Index.”). The HHI is calculated 16 by adding the squares of the individual firms’ market shares; it thus gives proportionately 17 greater weight to larger market shares. Trial Tr. 950:14-24 (Shapiro). The MERGER 18 GUIDELINES classify markets as highly concentrated if the HHI exceeds 2500. Trial Tr. 19 950:25-951:14 (Shapiro); MERGER GUIDELINES § 5.3. The MERGER GUIDELINES also look 20 at the increase in the HHI caused by the merger. “Mergers resulting in highly concentrated 21 markets that involve an increase in the HHI of more than 200 points will be presumed to be 22 likely to enhance market power.” MERGER GUIDELINES § 5.3; Trial Tr. 951:3-10 23 (Shapiro). 24 25 26 27 28 12 Rating-System’s 78 paying customers are primarily credit card swipe customers who pay very low fees for basic R&R, like the PowerReviews Express customers. Rating-System’s customers are not enterprise customers and Rating-System will not constrain Bazaarvoice’s pricing of R&R. Trial Tr. 979:2-12 (Shapiro). 68 1 166. Using shares based on IR500 customer revenues, the pre-merger HHI was 2674, 2 and it would increase by 1240 to 3915 after the merger. GX985* at Ex. 15A; GX1063; 3 Trial Tr. 955:22-956:7 (Shapiro). Based on IR500 customer count, the pre-merger HHI 4 was 2365, and it would rise by 2226 to 4590 after the merger. GX985* at Ex. 14A; 5 GX1062; Trial Tr. 950:25-951:11 (Shapiro). By either measure, these figures are 6 significantly above the MERGER GUIDELINES’ thresholds and create the presumption that 7 that the transaction would substantially reduce competition. See Trial Tr. 956:5-7 8 (Shapiro); MERGER GUIDELINES § 5.3. 9 167. Bazaarvoice complained that Dr. Shapiro has previously written that market shares, HHIs, and other market concentration measures have little relevance to unilateral effects 11 United States District Court Northern District of California 10 cases involving differentiated products, like this one. DX1384 at n.53; DX1736 at ¶¶ 178- 12 179; Trial Tr. 1034:4-7 (Shapiro). That is not quite accurate. Dr. Shapiro pointed out in 13 the article in the Antitrust Law Journal, The 2010 Horizontal Merger Guidelines: From 14 Hedgehog to Fox in Forty Years, from which Bazaarvoice quoted, that the market shares 15 of merging firms and the change of HHI are more informative than the level of HHI (at p. 16 720-721). And while the MERGER GUIDELINES reduce, they certainly do not eliminate, 17 emphasis on market shares. Id. The framework used by Dr. Shapiro for this case was 18 consistent with the analysis contained in his article. 19 168. In examining market concentration, Dr. Shapiro also considered supply 20 substitution. Supply substitution considers whether there are companies that are not selling 21 R&R platforms who could “very easily and quickly do so.” Trial Tr. 970:2-9 (Shapiro); 22 MERGER GUIDELINES § 5.2. Such firms are called “rapid entrants.” Trial Tr. 970:5-13 23 (Shapiro). The excess capacity a rapid entrant can divert to the new market must be 24 “measurable.” Trial Tr. 972:11-13 (Shapiro); MERGER GUIDELINES § 5.2. In such an 25 event, the rapid entrant would receive market share in accordance with its divertible 26 capacity. Trial Tr. 2068:1-11 (Shapiro). 27 28 169. Dr. Shapiro opined, and the Court concludes, that there are no rapid entrants in this case. First, the concept of divertible capacity does not apply in a software market like the 69 1 R&R platform market which is focused on developing a solution rather than converting 2 production from one product to another. As a result, there is no capacity-based market 3 share to assign to a rapid entrant. Trial Tr. 971:18-24, 2068:5-11 (Shapiro). Second, as 4 discussed later, firms outside the R&R platform market cannot easily and quickly enter the 5 market. To enter the R&R platform market and achieve some commercial success requires 6 significant investment and risk. Trial Tr. 972:7-10, 2068:12-22 (Shapiro). 7 170. Bazaarvoice attacks Dr. Shapiro’s market share analysis in numerous ways. It 8 complains that the IR500 does not capture huge swaths of the government’s alleged 9 relevant market and it excludes significant sectors of the economy, industries, and companies that use R&R. DX1398 at ¶ 5; DX1736 at ¶¶ 36-41, 71; DX1579; Trial Tr. 11 United States District Court Northern District of California 10 1537:9-1539:16; Trial Tr. 227:21-228:10. Bazaarvoice also asserts that the IR500’s focus 12 on retailers biases the market share calculation in favor of PowerReviews because 13 PowerReviews’s business strategy was to focus on retailers, Trial Tr. 1746:10-21 and the 14 IR500 does not capture R&R sales representing nearly 75 percent of Bazaarvoice’s 15 revenues and 34 of its top 50 customers, which are firms that are excluded from the IR500, 16 including many brands, DX1736 at ¶¶ 36, 37; Trial Tr. 1088:2-1089:11 (Shapiro). But Dr. 17 Shapiro showed that 72.7 percent of Bazaarvoice’s R&R revenue from brands (who are not 18 represented in the IR500) came from those that syndicate to retailers (many of whom are 19 presumably in the IR500). GX984 at 21; GX1059. Approximately 88 percent of its 20 revenue comes from retailers and manufacturers. Trial Tr. 273:20-274:10 (Hurt). While 21 significant third-party marketplaces that have built their own R&R solutions, such as eBay 22 and Amazon, are excluded from the IR500, they are captured in Dr. Shapiro’s analysis. 23 See, e.g., Finding of Fact 157 (Amazon), Trial Tr. 1082:23-1083:2 (Shapiro) (eBay). Each 24 check employed by Dr. Shapiro, including one using information from Dr. Shehadeh’s list 25 of customers, manifested the same basic market structure, so this argument fails. 26 171. Bazaarvoice complains that the IR500 provides no information about the many 27 thousands of eCommerce retailer sites that use R&R products supplied by other vendors, 28 such as Yotpo, Ratings System, Practical Data, Magento, and PowerReviews Express. 70 1 DX1579. While all of these solutions and the customers that use them are in the relevant 2 R&R platform market, they only have minimal market shares because they primarily sell 3 R&R at the lowest end of the market, which, in any event, Dr. Shapiro gauged in his 4 review of the IR1000 and the sample from the companies Dr. Shehadeh identified. This 5 criticism is unfounded. 172. 6 Bazaarvoice attacks the accuracy of the IR500 data because the data are collected by Internet Retailer by mailing a simple, unverified survey to retailers appearing in the 8 prior years’ IR500 list and retailers that may have “come to the attention” of Internet 9 Retailer. DX1398 at ¶ 6. It also points out that retailers’ online revenue is not necessarily 10 correlated with the value a retailer receives from R&R. Social commerce tools, including 11 United States District Court Northern District of California 7 R&R, may influence consumer behavior both online and offline at brick and mortar stores. 12 DX1900 (Dellarocca Dep. 140:1-141:4); GX986 at 4; DX1432 at 8. The IR500 data has 13 deficiencies, to be sure. Jason Goldberg, Bazaarvoice’s proffered industry expert, noted 14 both that he uses the IR500 “frequently” in his daily work but that the data “have 15 significant limitations.” 13 DX1432 at 71. It is a well-recognized source of information in 16 the eCommerce market. And Dr. Shapiro’s robustness checks, including those performed 17 using the Fortune 500 and the Shehadeh-based sample data, showed consistent results that 18 the merged entity controls well over 50 percent of the market. Despite the problems 19 identified with the data, Dr. Shapiro’s use of it was appropriate in analyzing the market and 20 market shares. 173. 21 Bazaarvoice complains that Dr. Shapiro’s market share checks are unreliable: the 22 Fortune 500 does not focus on customers that would likely use R&R, Trial Tr. 975:6-8, 23 976:5-7; the IR1000 uses the same methodology as the IR500, DX1778 at ¶ 8, Trial Tr. 24 969:5-9; and when Dr. Shapiro calculated shares using Dr. Shehadeh’s sample, he found 25 that in-house supply was larger than either Bazaarvoice or PowerReviews, Trial Tr. 26 13 27 28 As discussed more fully in the Order on Evidentiary Objections, the Court determined that Mr. Goldberg had sufficient qualifications to testify generally about the overall eCommerce industry but lacked sufficient expertise to provide opinions about the R&R market or central issues in this case. His reports and opinions were deficient in several ways outlined in that Order. 71 1 980:19-981:9. But none of Dr. Shapiro’s methods were subject to manipulation and each 2 rendered the same essential conclusion regarding the market share of the combined entity. 3 To underscore this point, when Dr. Shehadeh’s preferred grouping was defined to include 4 only American companies that used R&R, the market shares derived were consistent (59 5 percent) with Dr. Shapiro’s analysis of the IR500 data (68 percent [customer count] and 56 6 percent [revenue]). 7 174. Bazaarvoice argues that the MERGER GUIDELINES (1992) state that rapid entrants are market participants and that in a bidding market where market participants are equally 9 likely to secure sales, shares should be allocated equally among them. Rapid entrants 10 should be considered market participants for purposes of calculating market shares and 11 United States District Court Northern District of California 8 market concentration. MERGER GUIDELINES (1992) § 1.32 (referring to rapid entrants as 12 “uncommitted entrants”); see also MERGER GUIDELINES § 5.1 (“Firms that clearly possess 13 the necessary assets to supply into the relevant market rapidly may also be rapid 14 entrants.”). Bazaarvoice urged that Amazon, Google, Salesforce, Facebook, and Oracle, 15 among others, have the requisite technology, reputation, and relationships with brands and 16 retailers to be viewed as reasonable rapid entrants into R&R such that entry could occur 17 within one year. Bazaarvoice also contends that, conservatively, there are at least five 18 existing, equally competitive R&R options that customers have chosen in bidding 19 situations—Bazaarvoice, internal solutions, Gigya, Pluck, and Reevoo. Bazaarvoice 20 concludes that because at least seven competitors today have an equal likelihood of 21 securing sales in R&R for any given customer—Bazaarvoice, Pluck, Gigya, Reevoo, 22 internal solutions, Amazon, and Google—each competitor has an equal market share of 23 14.8 percent. 24 175. The next section explains why this calculation bears no weight. The evidence 25 established that there are potential entrants but no rapid entrants on the horizon and that all 26 remaining R&R companies are substantially weaker than PowerReviews as shown by their 27 market share. The range of advantages possessed by Bazaarvoice make it far and away the 28 leading choice for R&R platforms. As Dr. Shapiro commented, the mere existence of 72 1 other providers is insufficient to mitigate the transaction’s anticompetitive effects: “it’s 2 really not the number of bidders; it’s their quality and attractiveness to customers that 3 matters.” Trial Tr. 916:12-917:2 (Shapiro). This is not the ideal bidding market suggested 4 by Bazaarvoice that would allow use of equal market shares because the products are 5 differentiated, the competitors are of vastly different quality and the margins are quite 6 large. Trial Tr. 1035:7-1037:5 (Shapiro). As the court wrote in Chicago Bridge & Iron 7 Co. N.V. v. F.T.C., 534 F.3d 410, 429-30 (5th Cir. 2008), “Courts have generally 8 concluded that for entry to constrain supracompetitive prices, the entry has to be of a 9 ‘sufficient scale’ adequate to constrain prices and break entry barriers.” That is not the case here. 10 United States District Court Northern District of California 11 B. Bazaarvoice cannot rebut the government’s prima facie case: entry or repositioning from other firms has not, and will not, counteract Bazaarvoice’s increased market power 12 13 i. The Strength of PowerReviews Pre-acquisition 14 176. 15 Bazaarvoice attempted to show that PowerReviews was a weak competitor that 16 would be easily replaced by other competitors. The Court is not persuaded. Immediately 17 prior to being acquired by Bazaarvoice, PowerReviews had achieved the highest quarterly 18 sales in its history. GX643; GX87 (Godfrey CID Dep. 181:21-182:5); GX88 (Hossain 19 Dep. 196:4-6, 196:9-18). As discussed earlier, PowerReviews continued to compete 20 aggressively with Bazaarvoice up until the time of the acquisition. GX89 (Hurt Dep. 21 360:19-361:14); see also GX1264 at -759 (PowerReviews’s CEO received suggestions to a 22 press release in April 2012 so that “the C-level at BV will now read this as PR very 23 squarely coming after them”); GX419 at -484; GX89 (Hurt Dep. 377:14-378:16); GX1105 24 at -047. Indeed, less than two months prior to the merger, PowerReviews announced that 25 it had “added more than 30 major brands to its roster.” 14 GX619 at -390. 26 27 14 28 In this context, the Court understands that PowerReviews used “brands” to refer to both retailers and manufacturers. 73 1 177. PowerReviews’s former CEO, Ken Comée, testified that he was “confident” that he 2 “could have built even more value” for PowerReviews shareholders had the company not 3 been acquired. GX84 (Comée Dep. 201:17-20, 202:3-4); Trial Tr. 1839:5-24 (Comée). 4 178. Bazaarvoice pointed to huge losses on PowerReviews’s financial statements in the 5 last two months of its operation as evidence that PowerReviews would have collapsed of 6 its own weight if Bazaarvoice had not purchased it. Dr. Shapiro stated that “there was 7 some accounting thing going on” since the statement does not reflect “normal operations.” 8 Trial Tr. 1047:3-9 (Shapiro). The reasons for these losses were not adequately addressed 9 for the Court to reach any conclusions about them, but given PowerReviews success in adding customers and increasing revenue, it is clear that PowerReviews continued to be a 11 United States District Court Northern District of California 10 strong competitive constraint on Bazaarvoice until its acquisition. 12 179. Prior to the merger, Bazaarvoice estimated that PowerReviews could continue 13 operating up to 18 months before seeking any type of additional cash investment. GX87 14 (Godfrey CID Dep. 182:8-183:3). Pacitti fully expected that PowerReviews “intended to 15 continue competing with Bazaarvoice if the company were not to be acquired by 16 Bazaarvoice.” GX492 (Pacitti Dep. 145:7-11). Although Luedtke testified that 17 PowerReviews would have run out of cash in the second half of 2012, if that was so there 18 were a number of sources to which the company could have turned for additional funding 19 other than Bazaarvoice. GX83 (Collins 30(b)(6) Dep. 60:5-61:4). That included the 20 potential acquisition of PowerReviews by companies other than Bazaarvoice about which 21 Hurt had worried. GX83 (Collins 30(b)(6) Dep. 78:15-79:3). 22 180. PowerReviews had approximately 80 customers in the IR500 (estimates varied 23 between 78-86) and more retail customers than Bazaarvoice when it was acquired. It was 24 unquestionably Bazaarvoice’s closest competitor. For the foreseeable future, were it not for 25 the acquisition, PowerReviews would have likely continued to compete aggressively with 26 Bazaarvoice to maintain and expand its customer base. 27 28 74 ii. 1 Other Providers of R&R Platforms Are Unlikely to Expand Sufficiently to Mitigate the Transaction’s Anticompetitive Effects 2 3 4 5 6 7 8 9 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 181. The Court must determine whether entry or expansion would be timely, likely, and sufficient to replace the competitive constraint PowerReviews would have provided absent the acquisition. Trial Tr. 1713:16-1714:11 (Shehadeh); MERGER GUIDELINES § 9. There are a number of providers of R&R platforms in the relevant market that compete with Bazaarvoice for customers, and Bazaarvoice treats them as competitors. Trial Tr. 1008:318 (Shapiro). But as Dr. Shapiro testified, “if these products offered by these fringe suppliers were almost as good as what PowerReviews is offering, [one] would expect their commercial success to be also close to that of PowerReviews,” Trial Tr. 2063:19-2064:1 (Shapiro). None has achieved any meaningful level of commercial success in providing R&R platforms to domestic customers. 182. Bazaarvoice emphasized the strength of its current competition and the ability of rapid entrants like Google and Amazon to come into the market. It noted that the number of firms offering R&R solutions today compared with two years ago demonstrates that the technology necessary to develop a commercially acceptable R&R product is widely available and entry is therefore easy. Demand for integrated platforms has caused eCommerce platform providers to add social commerce features to their platforms, including R&R, so that their customers do not have to purchase products and services from multiple third-party vendors. Demand for integrated solutions has also caused social commerce vendors, such as Pluck, Gigya, and Lithium, to add social commerce features like R&R to their platforms. 183. Bazaarvoice argues that the industry’s move towards competition in “bundles” or “suites” of social commerce products has and will continue to induce entry. In the recent past, firms that offer suites have added R&R functionality to their offering. This trend towards vendors offering more functionality may well continue. 184. Bazaarvoice argues that the testimony of social commerce vendors underscores that entry is objectively easy. It notes that one person developed Rating System’s initial R&R 75 1 software by working part-time on the code for six months and that 2 It identifies companies like Gigya, Pluck, 3 and Lithium as examples of rapid entrants that have extended their product lines to include 4 R&R without much investment in the last few years. But these and other companies’ 5 relative lack of success selling R&R undercuts this argument, as do the lack of historical 6 entry and entry barriers like syndication and switching costs. Trial Tr. 2067:11-2070:24 7 (Shapiro). 8 9 185. Dr. Shehadeh testified that there has been significant entry, expansion, and repositioning since the time of the merger. Trial Tr. 1779:25-1782:16 (Shehadeh); DX1433 at 27-68; DX1736 at 67. He opined that historic entry patterns are not 11 United States District Court Northern District of California 10 informative because they do not demonstrate what would happen in the event of a price 12 increase. Trial Tr. 1778:24-1779:24. He identified Amazon as an existing competitor 13 through Amazon.com, Amazon Marketplace, and Amazon Webstore, as well as a potential 14 rapid entrant into the commercial supply of R&R. Trial Tr. 1783:17-1787:25. He pointed 15 to the expansion of smaller suppliers like Yotpo and Rating Systems, Trial Tr. 2085:21- 16 2086:23, as well as the entry of international suppliers like Reevoo and eKomi, Trial Tr. 17 2019:17-23; DX1736 at ¶¶ 211-220. He concluded that customers of all sizes can go in- 18 house and any customer can credibly threaten to do so. Trial Tr. 1871:5-1874:14 19 (Shehadeh); DX1736 at 54-56; DX1897 at 38. And he pointed out that Dr. Shapiro agreed 20 that a competitor need not win business to provide a competitive constraint. Trial Tr. 21 1058:13-17 (Shapiro). But as the description of the competitors below shows, none of 22 Bazaarvoice’s or Dr. Shehadeh’s arguments is persuasive because the competition is either 23 too weak or is merely complementary to Bazaarvoice’s and PowerReviews’s product, not a 24 substitute. 25 186. Reevoo is a London-based R&R platform provider that has been successful in 26 Europe. It entered the United States R&R platform market in September 2012. GX73* 27 (Reevoo Dep. 8:10-19) (reviewed in camera); GX87 (Godfrey CID Dep. 210:18-23). It 28 also sells a Q&A solution. GX73* (Reevoo Dep. 8:23-25) (reviewed in camera). 76 1 Bazaarvoice has competed with Reevoo in a number of accounts, both in the United States 2 and abroad. Reevoo launched an initiative targeted at Bazaarvoice and legacy 3 PowerReviews customers called the “swap-out campaign,” which Bazaarvoice vigorously 4 fought against. DX719. DX692 at -866. Bazaarvoice alleges that it has had to make 5 concessions to current or potential customers in response to Reevoo’s competitive 6 presence. 7 8 187. While Reevoo’s product is competitive with PowerReviews, it has struggled in the United States. Since entering the United States, Reevoo has contacted over GX73* (Reevoo Dep. 32:22-33:9, 33:12-15, 34:8- 9 10, 34:13-24, 37:3-22 (reviewed in camera), 41:7-10 (reviewed in camera), 41:13-16 11 United States District Court Northern District of California 10 (reviewed in camera)). 12 13 188. On June 17, 2013, Collins testified that he was “not sure to what extent, if any, [Reevoo] have entered the U.S. market.” GX81 (Collins Dep. 218:7-17). Reevoo’s domestic sales force consists of 14 GX73* (Reevoo Dep. 13:20-14:4) (reviewed in 15 16 camera); see also GX947* at -403. . GX73* (Reevoo Dep. 17 7:23-25, 13:20-14:7) (reviewed in camera). GX73* (Reevoo Dep. 14:8-11). 18 19 . GX73* (Reevoo Dep. 21:25-22:5, 20 21 22:8-18, 39:16-40:4) (reviewed in camera); GX125 (Clorox Dep. 122:9-18); Trial Tr. 22 294:1-16 (Levin/Clorox). 23 24 25 190. . GX73* (Reevoo Dep. 15:2-7, 15:12-20:8, 20:12:-21:11) (15:2-7, 15:1214, 20:7-8, 20:13-21:08 reviewed in camera). 26 GX73* (Reevoo Dep. 20:7-21:11, 37:23-38:3, 38:6- 27 28 39:4) (reviewed in camera). 77 1 [ GX73* (Reevoo Dep. 25:13-16, 25:19-26:9) (20:7-8, 20:13- 2 21:8, 37:23-38:3, 38:6-39:4 reviewed in camera). 3 4 . GX73* (Reevoo 5 6 Dep. 26:10-13, 26:16-27:25) (reviewed in camera). Reevoo is also hindered by its lack of reputation in the United States. 7 8 GX947* at -403. 9 Home Depot agreed that Reevoo’s recent entry into the United States market “raised a 11 United States District Court Northern District of California 10 flag” as Reevoo did not have a “proven . . . track record with what [Home Depot] would 12 classify major enterprise customers.” GX146 (Home Depot Dep. 67:13-68:4). 13 193. Pluck, a division of Demand Media, sells an array of social commerce applications, 14 including a R&R platform. GX62* (Pluck Dep. 149:22-25) (reviewed in camera). Collins 15 testified that Pluck is Bazaarvoice’s closest remaining competitor. GX81 (Collins Dep. 16 221:7-17); Trial Tr. 758:21-759:4 (Collins); see also GX492 (Pacitti Dep. 141:8-12). 17 GX62* (Pluck Dep. 169:25-170:5) (reviewed in 18 19 camera). 20 (reviewed in camera). GX62* (Pluck Dep. 135:16-21, 136:2-4). 21 22 . GX62* (Pluck Dep. 182:6-10) 194. 23 24 . GX62* (Pluck Dep. 61:20-22, 62:2-3, 184:25-185:4, 25 185:7-11, 185:13-16, 185:18-187:9, 187:12-188:1) (reviewed in camera). Some of Pluck’s 26 customers acknowledge that Bazaarvoice’s R&R platform is more robust and offers better 27 moderation and customization capabilities than Pluck’s R&R platform. GX100 (AutoZone 28 Dep. 33:14-34:12, 57:11-18, 57:21-58:12, 58:15-59:1); Trial Tr. 1333:6-11 78 1 (Rutherford/AutoZone); 2 The company has the technological capability to syndicate reviews, GX62* (Pluck Dep. 46:14-22, 59:1-4) (reviewed in 3 4 camera). GX62* 5 6 (Pluck Dep. 58:10-12, 58:15-25, 59:5-14) (reviewed in camera). . GX62* (Pluck Dep. 7 8 9 65:10-66:3). 195. . GX62* (Pluck Dep. 170:18-23). In 2011, PowerReviews judged that Pluck was “terrible” and that Target 11 United States District Court Northern District of California 10 “made a very bad choice.” GX813 at -737. 12 13 GX62* (Pluck Dep. 48:12-25, 49:21-50:10, 171:1-10). Target recently 14 discontinued its relationship with Pluck and switched to Bazaarvoice. Trial Tr. 759:5-11 15 (Collins); Trial Tr. 436:21-24 (Luedtke). 16 196. Gigya, founded in 2007, provides a suite of eCommerce applications, including 17 R&R, for online businesses. GX64 (Gigya Dep. 81:18-82:9). Its R&R product can be 18 bought separately, or as part of the full suite. Gigya’s suite of services include a social log- 19 in tool for a single point of authentication and a bundle of competitively priced social 20 commerce tools. Bazaarvoice and Gigya compete, and Bazaarvoice has offered to adjust 21 the parameters of potential deals after learning that customers were considering Gigya. 22 Trial Tr. 611:19-611:21, 619:1-619:10, 622:22-622:24 (Tarkowski); DX0391. 23 GX461; GX64 (Gigya Dep. 72:12-15, 24 25 26 83:24-84:14); Trial Tr. 600:5-14 (Tarkowski/Gigya). 197. Gigya’s R&R application lacks a network, does not offer syndication, and lacks 27 other functionality that some customers require, such as moderation. GX131 (Dillard’s 28 Dep. 58:11-59:18); GX810 at -442; GX64 (Gigya Dep. 92:24-93:12, 93:14-94:5, 101:479 1 16); Trial Tr. 603:4-15 (Tarkowski/Gigya); GX196 (World Kitchen Dep. 62:6-64:3); Trial 2 Tr. 294:22-295:11 (Levin/Clorox). Gigya’s R&R application is a more basic application 3 that does not allow consumers to provide ratings on specific product attributes nor does it 4 allow extensive customization. Trial Tr. 1208:15-1209:7 (Katz/Pacific Sunwear). Gigya 5 also does not allow retailers to use the application on an eCommerce website without 6 displaying the Gigya name. GX131 (Dillard’s Dep. 58:11-59:18). 7 198. Many of Gigya’s eCommerce customers, including those that bought Gigya’s full 8 suite and therefore have access to Gigya’s R&R application, use Bazaarvoice’s R&R 9 platform instead. GX64 (Gigya Dep. 105:7-21, 23); GX465 at -632. At least until trial, Bazaarvoice believed that Gigya targets different customers than Bazaarvoice. GX87 11 United States District Court Northern District of California 10 (Godfrey CID Dep. 210:24-211:9). Because Gigya and Bazaarvoice share customers, 12 Gigya’s CEO sees Bazaarvoice “more as a partner” than a competitor. GX465 at -632; 13 Trial Tr. 601:1-603:3 (Tarkowski/Gigya). Bazaarvoice shared that view. GX364 at -612; 14 see also GX762 at -070. 15 199. Lithium, founded in 2001, sells a software platform that promotes online 16 interaction between brands and consumers. GX72* (Lithium Dep. 16:12-19, 17:21-25). 17 The company’s “strength” is in fostering brand/consumer interactions “after [the] point of 18 purchase.” GX72* (Lithium Dep. 18:1-19:17). Lithium’s platform includes a variety of 19 applications, including forums, blogs, and Q&A. GX72* (Lithium Dep. 19:22-20:7). 20 R&R functionality is not included, but is available for purchase by customers who buy 21 Lithium’s platform. GX72* (Lithium Dep. 24:21-25:11). Lithium does not sell its R&R 22 application on a standalone basis, i.e., separately from its platform. GX72* (Lithium Dep. 23 24:24-25:18). 24 200. . GX72* (Lithium Dep. 19:18-21). . GX72* (Lithium Dep. 25 26 42:3-6). 27 28 Schuster Tr. 67:5-9, 69:23-70:1, 70:7-14, 137:15-22, 138:6-139:8. 80 1 201. . 2 3 GX72* (Lithium Dep. 136:3-16). GX72* (Lithium Dep. 136:23-137:12, 67:22-25, 68:2-10). 4 5 GX72* (Lithium Dep. 46:5-16, 46:19-47:9, 6 7 8 47:13-17). 202. Lithium and Bazaarvoice provide complementary solutions. GX364 at -609. Many of Lithium’s customers use Bazaarvoice’s R&R platform. GX72* (Lithium Dep. 47:20- 10 48:5); see also GX668. Before trial, Bazaarvoice management said that it “almost never 11 United States District Court Northern District of California 9 compete[s]” with Lithium. GX289 at -162; see also GX763; GX668; GX259 at -652; 12 GX87 (Godfrey CID Dep. 211:15-212:4). Bazaarvoice’s Chairman testified that Lithium 13 is “probably” not a current competitor to Bazaarvoice. GX491 (Meredith Dep. 200:6- 14 201:5). 15 203. Practical Data: Practical Data has sold add-on functionality, including R&R, to 16 users of the Yahoo! Store eCommerce platform since 2002. GX65 (Practical Data Dep. 17 6:2-22). Practical Data has approximately 110 customers. GX65 (Practical Data Dep. 18 19 43:4-7, 67:20-68:4, 68:7-16, 68:18). GX985* at Ex. 14A; GX1062. Practical Data does not offer moderation or 20 21 syndication services in connection with R&R. GX65 (Practical Data Dep. 53:8-54:11). 22 Practical Data does not have a research and development program or a roadmap for its 23 R&R product. GX65 (Practical Data Dep. 54:12-17). 24 204. Rating-System.com: Rating-System has no employees. GX66 (Rating-System 25 Dep. 89:7-11); Trial Tr. 1325:2-6 (Bruksha/Rating-System). The company’s founder 26 works full-time at Men’s Wearhouse. Several contractors provide services part-time as 27 needed. GX66 (Rating-System Dep. 13:19-14:3, 89:12-90:1); Trial Tr. 1316:17-20, 28 1325:7-21 (Bruksha/Rating-System). Rating-System offers several levels of R&R 81 1 services, including an “enterprise” plan for larger customers with “bigger traffic sites.” 2 GX66 (Rating-system Dep. 45:12-22); Trial Tr. 1318:23-1319:4 (Bruksha/Rating-System). 3 The company’s founder does not know how many R&R customers Rating-System has 4 (Trial Tr. 1318:10-22, 1319:25-1321:7 (Bruksha/Rating-System), but he testified that no 5 one has bought Rating-System’s enterprise service plan. GX66 (Rating-System Dep. 6 45:23-46:2); Trial Tr. 1319:5-9 (Bruksha/Rating-System). Rating-System does not offer 7 syndication, and only one customer uses its moderation services. GX66 (Rating-System 8 Dep. 120:14-121:15); Trial Tr. 1326:8-10 (Bruksha/Rating-System). Godfrey described 9 Rating-System as a “very, very small entrant into the market.” Trial Tr. 1129:22-1130:1 10 United States District Court Northern District of California 11 (Godfrey). 205. In 2008-2009, Viewpoints (now Wavetable Labs) built a R&R platform for Sears, 12 which replaced Bazaarvoice’s R&R platform on some of Sears’ websites. GX75 (Moog 13 Dep. 25:7-19, 31:2-17). After licensing its R&R platform to Sears, Viewpoints decided to 14 offer it to other customers. GX75 (Moog Dep. 31:2-17). Viewpoints made 100 sales calls 15 but signed only three additional customers. GX75 (Moog Dep. 35:20-37:12, 47:5-14). 16 None of those customers uses Viewpoints’ R&R platform today. GX75 (Moog Dep. 48:3- 17 24, 49:3, 49:5-7, 49:11-13, 49:19-22, 50:2-5). 18 206. Sears abandoned Viewpoints’ R&R platform in 2012 due to its cost, limited 19 functionality, and poor performance, leaving Viewpoints’ R&R platform with no 20 customers. GX181 (Sears Dep. 49:16-18, 49:20-50:3, 50:10-51:10); Trial Tr. 1690:16-25; 21 1691:9-25 (Massuda/Sears); GX75 (Moog Dep. 50:2-5). Viewpoints does not offer a R&R 22 platform today. GX75 (Moog Dep. 67:23-24, 68:3-17, 68:20). 23 207. Viewpoints believed that competing against Bazaarvoice for R&R customers would 24 have been a “bloodbath” because Bazaarvoice was a “very well funded and aggressive 25 company” with “lots of momentum and success” that “had proven to be very aggressive 26 about defending any opportunity.” GX75 (Moog Dep. 47:15-48:2, 69:16-20, 69:24-25, 27 70:3, 70:7-13). 28 82 208. 1 Yotpo is an Israeli company founded in or around the beginning of 2011. Its R&R 2 solution focuses on small to medium businesses. It sells its product throughout the world. 3 DX 1400 (Togrin Decl. ¶¶ 2, 10 and 11)). While Yotpo’s declarant claims that half of its 4 3000 R&R sales have been to US customers, none of those customers showed up in the 5 IR1000 or in Dr. Shehadah’s workable accounts sample reviewed by Dr. Shapiro. Trial Tr. 6 1065:19- 1066:7 (Shapiro). If Yotpo was more than a minor US competitor, the Court 7 assumes that more evidence would have been provided by Bazaarvoice about it. Because 8 the only information on Yotpo at trial came from a two page declaration, and the declarant 9 was not subjected to cross-examination, the Court gleans little about Yotpo other than that it is apparently competing in the United States for customers akin to those who use 11 United States District Court Northern District of California 10 PowerReviews Express. 15 12 iii. 13 In-House Solutions and eCommerce R&R Modules are Not Economically Viable Alternatives for Many Customers and Potential Customers of R&R Platforms 14 209. 15 Next to Bazaarvoice, companies that utilize in-house R&R platforms represent the second biggest market share for R&R platforms. This is an option for customers that 16 are not interested in the full feature set offered by Bazaarvoice (including moderation and 17 syndication services), customers that are willing to invest heavily in ongoing platform 18 development to maintain the software and create new features, or customers with specific 19 needs that no vendor meets. GX87 (Godfrey CID Dep. 213:8-23); GX661; GX195 20 (Wine.com Dep. 40:20-41:12); GX170 (Overstock.com Dep. 30:18-31:10, 32:23-33:10, 21 34:11-15, 68:3-12, 68:23-69:18, 72:7-9, 72:12-73:25). Some twenty-one of the one 22 hundred customer deponents had implemented an in-house solution, and twenty others said 23 that they would consider it. (Bazaarvoice listed these companies in its Post-Trial Proposed 24 Findings of Fact 247 and 248 (Dkt. No. 239), which the Court will not reproduce here.) 25 26 27 28 15 The parties’ pre-trial Stipulation arguably provided for admission of the Yotpo declaration in evidence. As discussed more fully in the Order on Evidentiary Objections, the Court admits the declaration over the government’s objection, but assigns it very little weight. 83 1 210. Customers who elect to build internal R&R platforms are “typically not” larger 2 enterprise customers looking for the type of sophisticated platforms provided by 3 Bazaarvoice and PowerReviews, but there are exceptions. GX87 (Godfrey CID Dep. 4 212:23-213:1). It is very expensive, in many cases prohibitively so, for customers to 5 develop an in-house R&R platform with functionality comparable to the features offered 6 by Bazaarvoice, and it would be difficult to maintain the same pace of innovation as 7 Bazaarvoice offers. GX272 at -217; GX625 at -356; GX885 at -572; GX781 at -634; 8 GX778 at -211-212; GX365* at -475-476, -478; GX928 at -929; GX939 at -551; GX160 9 (Lord & Taylor Dep. 38:9-21) cf. Trial Tr. 1690:4-15, 1692:5-13 (Massuda/Sears). Luedtke acknowledged that developing an in-house R&R platform was generally not a 11 United States District Court Northern District of California 10 good use of a potential customer’s limited internal technical resources. GX90 (Luedtke 12 Dep. 76:21-24, 77:3-10). 13 211. The cost of building a suitable in-house solution varies with the customer’s needs. 14 One PowerReviews customer, evaluating its options to build its own platform, estimated 15 that it would cost $120,000 up front and $12,000 per year in maintenance to replace the 16 R&R platform that it received from PowerReviews. GX168 (Onlineshoes.com Dep. 21:7- 17 22:6). A Bazaarvoice customer estimated it would cost $30,000-$40,000 for it to build its 18 own platform and then $35,000-$45,000 per year in maintenance and moderation costs. 19 GX114 (Build.com Dep. 17:17-19:7, 60:25-61:13); see also Trial Tr. 1190:20-1191:20 20 (Friedland/Build.com). Sears spent eleven months and approximately $2 million 21 developing an in-house R&R platform that does not have “a lot of the functionality” that 22 Sears would like to have. Trial Tr. 1689:20-1690:15, 1692:5-13 (Massuda/Sears). 23 212. The cost of performing moderation also discourages customers from choosing to 24 build an in-house platform. Godfrey called it an “extremely high barrier [to] entry.” 25 GX87 (Godfrey CID Dep. 213:8-13); Trial Tr. 1143:21-1145:16 (Godfrey); see also GX90 26 (Luedtke Dep. 81:21-82:21); GX193 (Walgreens Dep. 46:8-11, 46:14-16, 46:18-49:8); 27 GX364 at -603. It can be expensive for a customer to perform the same level of moderation 28 as is performed by third party providers. GX272 at -216; GX885 at -572; GX939 at -551; 84 1 GX89 (Hurt Dep. 204:8-20); Trial Tr. 106:18-107:24 (Hurt); GX125 (Clorox Dep. 146:22- 2 147:20). In its disclosures to the SEC, Bazaarvoice acknowledged that it benefits from 3 “economies of scale” in supplying moderation of R&R. GX964 at 79. 4 213. For customers who want syndication, building an in-house solution is generally not 5 a viable option. GX96* (Acer Dep. 71:2-12, 71:15-22); see also GX181 (Sears Dep. 6 51:17-18, 51:20-53:2, 53:4-23); Trial Tr. 1692:14-1694:4 (Massuda/Sears); GX125 7 (Clorox Dep. 146:22-147:20). Even if a customer built syndication technology, it would 8 not have access to a syndication network on which it could rely to receive and send R&R 9 content. GX142 (Green Mountain Coffee Dep. 55:11-17, 55:23-56:25). 10 214. There are companies such as Webcollage that syndicate content from brands to United States District Court Northern District of California 11 retail websites where the brand’s products are sold. Trial Tr. 1229:8-12 12 (Matthews/Webcollage). However, those syndication companies do not offer R&R 13 functionality. See, e.g., Trial Tr. 1239:11-17 (Matthews/Webcollage). Webcollage, in 14 fact, syndicates ratings and reviews for only “three or four” customers and each of those 15 customers have syndicated fewer than 50 reviews each. Trial Tr. 1239:23-25, 1240:1-5 16 (Matthews/Webcollage). 17 215. Many retailers have limited in-house developer resources, and building R&R 18 functionality, even if technically possible, just doesn’t “make a lot of sense” and is not a 19 “good use of development time.” GX138* (Fruit of the Loom Dep. 24:24-26:8); Trial Tr. 20 1227:9-15 (Krebs/Fruit of the Loom); see also GX144 (Hayneedle Dep. 46:6-47:4); Trial 21 Tr. 1638:5-15 (Moen/Hayneedle); GX141 (Golfsmith Dep. 32:12-14, 32:16-33:7); Trial 22 Tr. 1218:7-13 (Maki/Golfsmith); GX193 (Walgreens Dep. 46:8-11, 46:14-16, 46:18-49:8); 23 GX159 (LL Bean Dep. 37:6-38:16). One company described why designing and 24 maintaining an in-house solution would have been “too much work:” it would involve “UI 25 (user interface) design, graphic design, database design, building, testing and then 26 maintaining the system.” Trial Tr. 1352:24-1353:18 (Hughes/Astral Brands). 27 28 216. For the reasons above, some customers that have invested in building their own R&R platform “ultimately decided to scrap it and use Bazaarvoice.” GX885 at -572; see 85 1 also GX88 (Hossain Dep. 138:4-140:15). Generally, the adoption of commercial R&R 2 platforms, such as those offered by Bazaarvoice and PowerReviews, has been increasing 3 over time, while the adoption of in-house R&R platforms has been relatively steady. 4 GX983* at 28-29; GX1038; Trial Tr. 927:18-928:22 (Shapiro); GX1038. 16 217. 5 eCommerce platforms are a technology that enables a retailer to create an 6 eCommerce site. For example, eCommerce platforms provide the shopping cart 7 functionality on a retailer’s website. GX90 (Luedtke Dep. 45:21-46:5, 46:8-16). 218. 8 While some eCommerce platforms may offer some basic R&R functionality, they generally lack the features necessary to serve as a substitute for the R&R platforms offered 9 by Bazaarvoice or PowerReviews. GX196 (World Kitchen Dep. 19:8-14, 64:17-65:17); 11 United States District Court Northern District of California 10 GX90 (Luedtke Dep. 116:2-12, 116:15-20); Trial Tr. 419:6-13 (Luedtke); GX655 at -917, 12 -920. According to Bazaarvoice, “[m]ost e-commerce platforms have R&R widget[s] 13 only” that do not have the features available on the Bazaarvoice or PowerReviews R&R 14 platforms such as “reporting, robust analytics, content moderation, best practices for 15 gathering and using content, integrations with other systems, etc.” GX885 at -573; see 16 also GX364; GX68 (ShopVisible Dep. 45:23-46:2, 46:5-12, 46:15, 46:19-21, 46:23, 48:20- 17 23, 49:1-2, 52:4-8, 52:10-16, 52:18). Bazaarvoice’s “technology is more feature rich and 18 has a lot more bells and whistles” than other eCommerce platforms. GX81 (Collins Dep. 19 240:8-13); Trial Tr. 883:7-15 (Collins). Some customers may start by using native R&R 20 functionality in an eCommerce platform, but as they grow they reach the limits of the 21 platform’s R&R capabilities. GX189 (Tempur-Pedic Dep. 21:14-22:6); GX141 (Golfsmith 22 Dep. 60:25-61:10, 61:12-17). 219. 23 To the extent that some R&R “plug-in” modules exist, the modules require significant customization and development time by the retailer and as such are not 24 25 16 26 27 28 Bazaarvoice offered Exhibit 3a to Dr. Shehadeh’s rebuttal report to argue that the adoption of in-house platforms has increased substantially. This exhibit was based on an error Dr. Shapiro initially made concerning Amazon months before that materially over-stated the market in a way that supported Bazaarvoice’s argument. Dr. Shehadeh was well aware of that error, which had been corrected prior to trial, but inexplicably used it at trial anyway. His justification, that he was merely using Dr. Shapiro’s work to make his point, fell flat. Trial Tr. 1927:22-1928:2 (Shehadeh). 86 1 particularly good substitutes for a robust R&R platform. GX99 (Astral Brands Dep. 127:6- 2 12, 127:15-129:5); Trial Tr. 1351:20-1352:19 (Hughes). 3 220. Regardless of whether an eCommerce platform offers its own R&R functionality, 4 many also offer “app stores” that allow customers to purchase add-on R&R functionality 5 from other vendors, such as PowerReviews’s Express R&R product. These add-ons 6 improve upon the limited functionality available natively in the eCommerce platform. 7 GX90 (Luedtke Dep. 45:2-20, 46:11-16, 116:2-12, 116:15-20). 8 9 221. eCommerce platforms do not often compete head-to-head with Bazaarvoice or PowerReviews to offer R&R platforms. GX68 (ShopVisible Dep. 43:22-24, 44:2-5, 44:7); see, e.g., Trial Tr. 884:13-15 (Collins). Some eCommerce platforms purchase R&R 11 United States District Court Northern District of California 10 platforms from Bazaarvoice for their own websites. See, e.g., Trial Tr. 883:16-884:3 12 (Collins). Switching eCommerce platforms can be an expensive and complicated process 13 and customers generally won’t switch eCommerce platforms in order to obtain R&R 14 functionality. See, e.g., DX1418 and Trial Tr. 1226:18-1227:1 (Krebs/Fruit of the Loom); 15 Trial Tr. 1667:3-1668:3, 1669:13-1670:16 (Yukel/Big Dot of Happiness); Trial Tr. 16 1954:13-19 (Shehadeh). 17 iv. 18 19 222. Other Social Commerce Companies and Large Software Companies are Unlikely to Enter the Market for R&R Platforms Bazaarvoice says that it is actively concerned about the possibility of the “launch of 20 an R&R product by one of the big guys” like Oracle or Salesforce, and has had “strategic 21 conversations” on how to remain competitive if such a company launched a R&R product. 22 DX202 at -469. Collins noted that there is “another universe of startups and established 23 companies” that the company had to consider competitors, including Google, Facebook, 24 eBay, Yahoo, Merchantry, Steelhouse, and HookLogic. Bazaarvoice identified each of 25 these companies as “dangerous weapons if combined with other capabilities.” DX202 at - 26 469-470. As discussed below, no evidence was presented that any of these companies will 27 enter the market within the next two years. 28 87 1 223. TurnTo provides a social question and answer software system to online retailers. 2 While TurnTo’s CEO said that it is considering whether to develop a R&R solution and 3 believed it would take them “less than a year” to develop and release an R&R product, he 4 believes it would take many years to develop R&R functionality comparable to the 5 Bazaarvoice and PowerReviews platforms. GX74* (TurnTo Dep. 41:25-42:13, 173:16-20, 6 173:23-174:13,188:4-23). He believes that Bazaarvoice’s “strong” network makes it 7 “much harder” for a new entrant to sell R&R. GX74* (TurnTo Dep. 50:19-51:2, 51:5-25). 8 If TurnTo were to try to enter the R&R platform market, the CEO’s “fear” is that TurnTo 9 would enter with a less mature and less expensive R&R platform than Bazaarvoice offers. 10 United States District Court Northern District of California 11 GX74* (TurnTo Dep. 42:25-43:4, 43:7-44:17). 224. Amazon has the resources and skill to expand its R&R offering if it chose to. It 12 has demonstrated its willingness to enter bourgeoning industries. DX1735 at 32; Trial Tr. 13 661:7-662:1 (Barton), 1625:18-1626:1 (Ahmed). Bazaarvoice noted that Amazon has 14 “hundreds of millions of visitors,” “a lot of data” and “a lot of capabilities.” Trial Tr. 15 791:20-792:3 (Collins). Bazaarvoice asserts that nothing prevents Amazon from readily 16 entering the R&R business in competition with Bazaarvoice and other commercial vendors 17 simply by offering a license to its pre-existing “best in class” R&R technology. GX70 18 (Ahmed Dep. 89:1-3). It has the technology and could “unbundle” its R&R solutions and 19 offer them to customers. Trial Tr. 1783:17-1784:16 (Shehadeh). It is the leading 20 eCommerce retailer. GX71 (Sell Dep. 37:6-14). Retailers view Amazon as a competitive 21 threat; nevertheless, most admit they distribute their products through or partner with 22 Amazon and derive benefit from that relationship. Trial Tr. 1166:14-1167:12 (Friedland). 23 225. Amazon has strong relationships with both brands and retailers. Trial Tr. 1784:20- 24 1785:10 (Shehadeh). It could syndicate content. Trial Tr. 1784:20-1785:10 (Shehadeh), 25 1412:25-1413:5 (Meredith). It previously syndicated reviews from its R&R to external 26 eCommerce sites; and Amazon currently syndicates its reviews to subsidiaries, including 27 Quidsi, and to a “very large list” of “thousands” of Amazon “associates.” GX70 (Ahmed 28 Dep. 28:13-19, 82:9-18). 88 1 226. All of that said, there is no persuasive evidence that Amazon has plans to enter the R&R market commercially. Amazon “do[es] not compete with [Bazaarvoice] as their 3 platform does not allow third-parties to access review data for any reason.” GX672 at - 4 846. Further, Alan Godfrey, Bazaarvoice’s General Manager of Retail, testified that he 5 cannot recall lowering the price floor for R&R platforms in response to competition from 6 Amazon. GX1119 (Godfrey Dep. 148:7-10); see also Trial Tr. 1844:22-1845:5 (Comée) 7 (not aware of an instance of PowerReviews competing directly against an Amazon 8 standalone R&R platform). While Bazaarvoice points to the testimony from Waqas 9 Ahmed, Amazon’s Technical Product and Program Manager, that Amazon “almost daily” 10 considers entry into the commercial supply of R&R, Trial Tr. 1625:18-1626:1, in context 11 United States District Court Northern District of California 2 it appeared that Ahmed was merely saying that Amazon will always consider a business 12 opportunity. He gave no testimony that Amazon had done a preliminary analysis or taken 13 any steps to investigate market entry, and the Court does not conclude that Amazon should 14 be categorized as a rapid entrant. 15 227. While Amazon.com has had an in-house R&R platform for its own website for 16 over 18 years, Amazon.com does not currently provide its R&R platform to any website 17 that is not an Amazon subsidiary and has no plans to license its R&R platform to other 18 eCommerce vendors. GX70 (Amazon.com (Ahmed) Dep. 37:11-15, 114:11-24); Trial Tr. 19 1625:18-1626:6 (Ahmed/Amazon.com). 20 228. Amazon is unlikely to enter the market for R&R platforms with a commercial R&R 21 product because, as a document used to prepare for the Bazaarvoice IPO roadshow 22 indicated, “it would actually be disadvantageous for them to allow brands or retailers to 23 display the content; Amazon’s first goal for the content is to attract consumers to their site 24 and keep them there for shopping, which relies on their exclusive access to the data.” 25 GX672 at -846. Where retailers or manufacturers sell products on both their own website 26 and on Amazon.com, Amazon.com does not permit syndication of its R&R between the 27 two sites. GX70 (Amazon.com (Ahmed) Dep. 113:20-114:8); Trial Tr. 1624:7-19 28 (Ahmed/Amazon.com). 89 1 229. Amazon also sells an eCommerce platform called Amazon Webstore. Amazon 2 sells the Amazon Webstore eCommerce suite to approximately 3 whom have access to the platform’s R&R functionality. GX71 (Sell Dep. 32:13-17, 36:9- 4 17). Customers range from “companies that could be as small as one person to large 5 companies.” Id. at 17:11-17. Bazaarvoice showed that Alternative Apparel switched from 6 Bazaarvoice to Amazon WebStore in 2012 and that Fruit of the Loom switched from 7 PowerReviews to Amazon Webstore for R&R functionality in August 2012. GX1038 8 (Fruit of the Loom/Krebs Dep. 31:11-24). 9 230. customers, all of Although the Amazon Webstore eCommerce platform includes some R&R functionality, Defossé stated, “I have yet to run into a deal where Amazon is a direct 11 United States District Court Northern District of California 10 competitor to our overall services in a deal where they are not already embedded. So, net 12 net, I would not be concerned with them as a direct competitor . . . .” GX19; see also Trial 13 Tr. 1150:16-1151:8 (Godfrey). In fact, they cooperate rather than compete: Amazon 14 allows Bazaarvoice to sell the Express product to its Amazon Webstore eCommerce 15 platform customers. GX20 at -649; Trial Tr. 550:22-551:11 (Defossé). The R&R 16 functionality available on Amazon Webstore is only a subset of that on Amazon.com, as 17 Amazon.com generally does not make available its features to the Amazon Webstore 18 business unit. GX71 (Amazon Webstore (Sell) Dep. 21:17-25, 29:19-30:1); Trial Tr. 19 1706:12-25, 1707:9-16 (Sell/Amazon Webstore). Similarly, the analytics features 20 available on Amazon Webstore do not leverage any of the analytics on Amazon.com and 21 reviews submitted to retailers using the Amazon Webstore eCommerce platform cannot go 22 through Amazon.com’s moderation process. GX71 (Amazon Webstore (Sell) Dep. 29:11- 23 18, 39:24-40:2); Trial Tr. 1707:1-8, 1707:17-19 (Sell/Amazon Webstore). Amazon has no 24 plans to improve the R&R functionality available via the Amazon Webstore eCommerce 25 platform. GX71 (Amazon Webstore (Sell) Dep. 42:03-42:07); Trial Tr. 1708:12-16 26 (Sell/Amazon Webstore). 27 28 231. Amazon does not syndicate reviews between Amazon.com and retailers using the Amazon Webstore eCommerce platform. GX71 (Amazon Webstore (Sell) Dep. 40:2190 1 41:12); Trial Tr. 1707:20-1708:11 (Sell/Amazon Webstore). Furthermore, Bazaarvoice 2 does not believe that Amazon will start to syndicate R&R to its Amazon Webstore 3 eCommerce platform customers. GX672 at -846. It is currently difficult to import or 4 export ratings and reviews to or from a website using Amazon Webstore. A very limited 5 amount of ratings and reviews content can be exported via an application programming 6 interface (API), while importing ratings and reviews would require the manual 7 resubmission of the reviews on the Amazon Webstore-run website. GX71 (Amazon 8 Webstore (Sell) Dep. 41:13-42:1, 45:12-18, 45:21-46:12, 46:14-47:7). 9 232. Bazaarvoice does not view Facebook as a competitor but rather “more a partner than a potential competitor.” GX928 at -928; Trial Tr. 889:5-21 (Collins); see also GX655 11 United States District Court Northern District of California 10 at -916; GX492 (Pacitti Dep. 44:8-13). According to the Q&As drafted in preparation for 12 Bazaarvoice’s IPO roadshow, Bazaarvoice “view[s] Facebook as another channel for us to 13 reach consumers; they are not competitive to our business model, we are actually very 14 different and complementary.” GX425 at -927, -928. Facebook told Bazaarvoice 15 executives that it has “no plans to build out Reviews capabilities.” GX509 at -869. 16 Bazaarvoice represented to the SEC and potential investors that Facebook “do[es] not 17 currently focus on our market.” GX964 at 85; see also Trial Tr. 641:6-9 (Barton). 18 233. No Bazaarvoice brand customers have indicated an interest in abandoning 19 Bazaarvoice to rely on Facebook services. GX116 (Curtin 30(b)(6) Dep. 12:11-16). 20 Bazaarvoice’s Vice President of Brands, Matt Curtin, testified that he has not received any 21 requests from Bazaarvoice sales people for authority to provide a discount to Bazaarvoice 22 customers due to competition from Facebook. GX116 (Curtin 30(b)(6) Dep. 13:11-15). 23 234. In an interrogatory response, Bazaarvoice listed Facebook as a partner, i.e., an 24 “organization[] with which Bazaarvoice is, at this time, strategically aligned and/or share 25 Bazaarvoice’s commitment to pursuing innovative solutions that help businesses better 26 engage with customers.” GX987 (Def.’s Resp. to Pl. Interrogs. 20 & 21 7:22-24, 8:4). 27 More specifically, “Bazaarvoice offers a social commerce application for Facebook users 28 that permits shoppers visiting the product page of a client retailer to automatically send all 91 1 of their product reviews and other content to their personal page on Facebook.com.” 2 GX987 (Def.’s Resp. to Pl. Interrogs. 20 & 21 11:12-15). 3 235. Google is not a competitive constraint on Bazaarvoice and Bazaarvoice views 4 Google as “more a partner than a potential competitor.” GX928 at -928; Trial Tr. 889:5-21 5 (Collins); see also GX1133 at -569; GX511 at -570. Pacitti testified that Google does 6 “[n]ot today” “compete[] in any market which Bazaarvoice competes.” GX492 (Pacitti 7 Dep. 48:17-49:12); see also Trial Tr. 641:6-11 (Barton); Trial Tr. 695:18-696:11 8 (Osborne). According to the Q&As drafted in preparation for Bazaarvoice’s IPO 9 roadshow, Google is “not competitive to our business model, we are actually very different and complementary.” GX425 at -927. Google does not offer R&R platforms to retailers. 11 United States District Court Northern District of California 10 GX81 (Collins Dep. 253:22-25). Bazaarvoice represented to the SEC and potential 12 investors that Google “do[es] not currently focus on our market.” GX964 at 85. 13 Bazaarvoice has developed a standardized integration with Google Analytics. GX987 14 (Def.’s Resp. to Pl. Interrogs. 20 & 21 8:24-26). 15 236. Bazaarvoice has never lost a R&R platform customer to any Google product. 16 GX82 (Barton Dep. 173:17-174:15); Trial Tr. 641:6-11 (Barton); GX1119 (Godfrey Dep. 17 34:19-24). Nor did PowerReviews view Google as a direct competitor. GX88 (Hossain 18 Dep. 97:20-98:5). In its SEC filings, Bazaarvoice stated that it does not yet face 19 competition from Google and that Google “do[es] not currently focus on [Bazaarvoice’s] 20 market.” GX969 at 36; GX970 at 8. 21 237. IBM is a partner of Bazaarvoice and Bazaarvoice does not believe that IBM is 22 “motivated to come to our marketplace.” GX87 (Godfrey CID Dep. 167:9-22). In fact, 23 Bazaarvoice has an out-of-the-box integration/connector for its R&R platform with the 24 IBM WebSphere Commerce eCommerce platform. GX987 (Def.’s Resp. to Pl. Interrogs. 25 20 & 21 9:2-3, 9:5). Bazaarvoice has also developed a standardized integration with IBM 26 Coremetrics. GX987 (Def.’s Resp. to Pl. Interrogs. 20 & 21 8:24-25, 8:27). 27 28 238. Similarly, Oracle and Salesforce.com are partners of Bazaarvoice, not competitors. Bazaarvoice has an out-of-the-box integration/connector for its R&R 92 1 platform with the Oracle Commerce (ATG) eCommerce platform. GX987 (Def.’s Resp. to 2 Pl. Interrogs. 20 & 21 9:2-4). Bazaarvoice also has a solution that integrates with Oracle 3 and Salesforce.com’s solutions. GX987 (Def.’s Resp. to Pl. Interrogs. 20 & 21 13:8-11, 4 13:13-14). 5 239. The companies just discussed have the size and strength to enter virtually any 6 technology market and become strong competitors. But there is no credible evidence that 7 any of them are considering entry into the R&R platform market in the United States. The 8 Court concludes that there are no rapid entrants. 9 v. There are significant barriers to entering the market for R&R platforms 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 240. Bazaarvoice asserted at trial that there were low or no barriers to entry to the R&R market because R&R software was tending towards commoditization, the proof of which was the number of new competitors that have entered the market in the past few years. But even before the acquisition of PowerReviews, Bazaarvoice’s competitive position was protected by substantial barriers to entry, a fact which Bazaarvoice used to herald. GX835 at -825. As a part of its “Key Messages for Initial Public Offering,” prepared in February 2012, Bazaarvoice stated that there were “[s]ignificant barriers to entry, [it] would be very difficult for a new company to enter our market organically or through M&A,” and that it “[w]ould take a competitor years to build a comparable solution, but then we already have the customers and the network.” GX835 at -825; see GX425 at -927 (“the barriers to entry are very high in our current market”). 241. Bazaarvoice’s syndication network is a barrier to entry in the market for R&R platforms. Bazaarvoice emphasized the size and scope of its syndication network in its efforts to market the company. GX413 at -924; GX750 at 15; Trial Tr. 125:10-126:1; 146:19-147:4 (Hurt); GX425 at -929, 931. Bazaarvoice’s syndication volumes have been increasing “30-50% year over year.” Trial Tr. 889:25-890:15 (Collins). Economic analysis confirms that syndication is becoming “significantly more important” for brands and retailers Trial Tr. 1014:15-18 (Shapiro), and 93 1 it would be “very hard” for a new entrant to overcome the syndication barrier to entry, 2 Trial Tr. 1017:18-1020:20 (Shapiro). See also Trial Tr. 1014:19-1015:6 (Shapiro) and 3 GX1052 (growing adoption of syndication); Trial Tr. 1015:7-20 (Shapiro) and GX1053 4 (same); Trial Tr. 1016:1-1017:16 (Shapiro) and GX1059 (same). Dr. Shehadeh anticipated 5 that there are “trends up” in terms of the percentage of customers choosing syndication and 6 that he expects the number of syndication customers will continue to grow. Trial Tr. 7 1992:13-1993:10 (Shehadeh). 8 9 242. Bazaarvoice recognized that the acquisition of PowerReviews increases this barrier to entry, as it noted in the title to a PowerPoint slide, “Scale and Network Effects Create Sustainable Competitive Advantage.” GX770 at -232; Trial Tr. 145:16-25 (Hurt); see also 11 United States District Court Northern District of California 10 GX925 at -941-46; GX409 at -568. In the due diligence memo prepared by company 12 executives recommending the acquisition to the Bazaarvoice Board, Bazaarvoice 13 executives wrote that by virtue of the merger, “Bazaarvoice blocks market entry by 14 competitors and therefore we ‘cover our flank’ to ensure our retail business is protected 15 from direct competition and premature price erosion.” GX925 at -941; see also GX254 at 16 -340. A PowerReviews Board member noted that “post this combination, there is a 17 network effect between the brands and the retailers that will be nearly impossible for 18 someone to break.” GX746 at -897. A Bazaarvoice Board member similarly noted that the 19 acquisition “was a good one given the increased network power it would give 20 [Bazaarvoice] on the retail side.” GX949. 21 243. As more manufacturers purchase Bazaarvoice’s R&R platform, the Bazaarvoice 22 network becomes more valuable to retailers because it allows them to gain access to a 23 greater volume of R&R. Similarly, as more retailers purchase Bazaarvoice’s R&R 24 platform, the Bazaarvoice network becomes more valuable for manufacturers because it 25 will allow them to syndicate content to a greater number of retail outlets. The feedback 26 between manufacturers and retailers creates a network effect that is a significant and 27 durable competitive advantage for Bazaarvoice. GX770 at -219 and Trial Tr. 145:16-25 28 (Hurt); GX406 at -202 and Trial Tr. 124:15-125:2 (Hurt); GX983* at 73-76; see also 94 1 GX215 at -652; GX1119 (Godfrey Dep. 115:15-116:20); see also GX44 at -438; Trial Tr. 2 296:22-297:6 (Levin/Clorox). In its registration statement and accompanying prospectus 3 for its IPO, Bazaarvoice stated that “[o]ur ability to syndicate content across a wide array 4 of websites attracts brands to our network . . . . As a result, we believe we benefit from 5 powerful network effects that differentiate us from our competitors.” GX964 at 4, 79. 6 244. Bazaarvoice acknowledged the importance of its syndication network as a 7 substantial barrier to entry that protects its dominant position. Collins explained that its 8 “[d]ata and network effect should be a buffer against price deflation/commoditization as 9 long as we maintain significant market share and lock competitors out of a meaningful data set.” GX1228 at -319. Pre-merger, Osborne wrote to Hurt in January 2012: “We’ve seen 11 United States District Court Northern District of California 10 the network effect already in our renewals - customers that are using syndication generally 12 can’t leave us, and that’s a great thing.” GX1220 at -614 and Trial Tr. 711:1-712:6 13 (Osborne). Bazaarvoice prepared a document for an investor roadshow before its IPO in 14 which it explained the “powerful network economies at play by linking retailers and 15 brands/manufacturers.” GX425 at -919. Bazaarvoice identified its “ability to leverage the 16 data” from its customer base as “a key barrier [to] entry.” GX425 at -926. The company 17 claimed that “any company entering the market would have to start from the beginning by 18 securing all of the retail clients,” which would be difficult because most of the largest 19 retailers are already using the Bazaarvoice platform. GX425 at -926 and Trial Tr. 131:9-20 20 (Hurt). It boasted to potential investors that “the power of [Bazaarvoice’s] network effect 21 and significant advantage on a global scale is starting to crowd out competition . . . .” 22 GX425 at -928. 23 245. Since the IPO, Bazaarvoice’s SEC filings have continued to identify “powerful 24 network effects” from syndication as a “competitive strength[] [that] differentiate[s] 25 [Bazaarvoice] from . . . competitors” and “serve[s] as [a] barrier to entry.” GX501 at 7; 26 GX964 at 4, 79; see also GX87 (Godfrey CID Dep. 110:3-10); Trial Tr. 258:22-259:11 27 (Hurt). In explaining Bazaarvoice’s network effect, Curtin stated that having a majority of 28 the companies on the IR500 list as customers makes the Bazaarvoice network more 95 1 attractive to brands and that the “most important thing to a brand is the ability to take that 2 [R&R] content and syndicate [it] . . . .” GX1009 (Curtin Dep. 52:15-54:22, 96:21-98:15); 3 see also GX93 (Svatek Dep. 93:4-9). 4 246. PowerReviews also recognized the significant network effect of Bazaarvoice’s large syndication network, and acknowledged losing competitive opportunities because of 6 the strength of that network effect. GX90 (Luedtke Dep. 211:9-12, 211:22, 211:25-212:11, 7 235:7, 235:9-12, 237:18-24, 238:16-239:3, 242:3-12); Trial Tr. 423:6-25 (Luedtke); see 8 also GX245 at -433-34; GX410 at -482-83; GX89 (Hurt Dep. 232:7-24); GX669 at -630; 9 GX280 at -600. PowerReviews noted after a call with Clorox, a Bazaarvoice customer, 10 that “[i]f we could break the syndication barriers we could make significant inroads into 11 United States District Court Northern District of California 5 existing [Bazaarvoice] US retail accounts.” GX607* at -086. 12 247. The acquisition of PowerReviews will extend the reach of Bazaarvoice’s network 13 and deprive its remaining competitors of the scale that is necessary to compete effectively. 14 See GX87 (Godfrey CID Dep. 82:15-83:11); GX425 at -919, -923, -926, -933, -935; 15 GX313 at -633; GX899 at -816-17; see also Trial Tr. 128:15-129:18 (Hurt). Collins 16 predicted that Bazaarvoice’s acquisition of PowerReviews “would definitely tip the scales 17 in [Bazaarvoice’s] permanent favor on the network front.” GX522 at -035. Similarly, 18 Hurt wrote, in preparing for a Board meeting prior to the acquisition, “[o]ur syndication 19 network is a competitive advantage today, but it can become a true competitive moat in the 20 future.” GX411 at -871; see also Trial Tr. 262:4-11 (Hurt). Bazaarvoice anticipated that 21 the assimilation of major PowerReviews retailers into the Bazaarvoice network “further 22 increases the switching costs, and therefore deepens [Bazaarvoice’s] protective moat, for 23 brands and retailers alike.” GX925 at -943. Svatek testified that “the network . . . would be 24 very difficult for any individual company to replicate, any individual customer, a retailer or 25 brand to replicate.” GX93 (Svatek Dep. 143:18-144:2). 26 27 28 248. Other R&R platform providers acknowledge that syndication is very important and that Bazaarvoice’s syndication network gives it a competitive advantage. (reviewed in camera); GX72* (Lithium Dep. 39:12-40:4, 40:10-11, 96 1 45:7-46:16, 46:19-47:9, 47:13-17, 47:20-49:3, 49:11, 49:15-16, 49:18-50:4); (reviewed in camera). 2 3 249. At trial, Bazaarvoice tried to walk away from the importance of its syndication network. Dr. Shehadeh opined that syndication does not create network effects rising to 5 the level of a barrier to entry because most customers do not use it and it is not an essential 6 component of the R&R product. Trial Tr. 1765:24-1775:3, 2018:24-2019:16 (Shehadeh); 7 DX1736 at ¶¶ 146-152; DX1433 at ¶¶ 206-235; DX1897 at 25. Bazaarvoice offered 8 evidence both that syndication was valuable and evidence that syndication is not always 9 beneficial for eCommerce sites. Its proffered industry expert, Mr. Goldberg, testified that 10 syndication could cause a company’s search engine optimization effectiveness to decline 11 United States District Court Northern District of California 4 because of “duplicated” content. Bazaarvoice states that it “never had a client notify us of 12 a duplicate content warning in Google Webmaster Tools because of content syndication to 13 other sites, sharing within product families, or repurposing review content to be used for 14 another relevant purpose in the same site.” GX1248, p.1. Mr. Goldberg criticized 15 Bazaarvoice’s website as “misleading” by downplaying the problems created by its 16 syndication program. Trial Tr. 1593:15-17 (Goldberg). 17 250. Mr. Golderg also testified that eCommerce sites may not want syndication from a 18 manufacturer that has different standards for moderation or does not collect the same 19 information from its reviewers that the retailer collects. Trial Tr. 1527:3-11. Syndication 20 may not fit some customers’ business models or needs, especially retailers. Trial Tr. 21 442:17-21 (Luedtke). Some sites may have sufficient reviews not to need additional 22 reviews provided by other sites. Trial Tr. 1170:4-24 (Friedland). This is true especially of 23 retailers who have sufficient content generated on their own sites. GX170 24 (Overstock.com/Jolitz Dep. 36:18-37:1). And a brand might not distribute any goods 25 through any retailer and therefore would have no desire to syndicate reviews to any other 26 retailer’s website. GX94 (Abercrombie & Fitch/May Dep. 44:19-45:14); GX158 (La-Z- 27 Boy/Targett Dep. 22:23-23:3). 28 97 1 251. Moreover, it is possible for customers of non-Bazaarvoice R&R solutions, who 2 own the rights to their ratings and reviews, to syndicate their content even if they choose 3 an alternative R&R solution and share them with whatever sites they choose, including 4 non-Bazaarvoice customers. Trial Tr. 1605:18-25, 672:17-674:2; DX252 at -830. 5 252. Even if it is true that the majority of retailers and manufacturers today do not use 6 syndication, the trend is decidedly in favor of it, particularly for those in the market for 7 R&R platforms in the United States. Collins said that Bazaarvoice’s syndication 8 customers are growing 30-50 percent year over year. Dr. Shehadeh and Dr. Shapiro 9 differed over whether 27 percent or 37 percent of Bazaarvoice customers used syndication, but Dr. Shapiro showed that 72.7 percent of Bazaarvoice’s revenue from manufacturers, 11 United States District Court Northern District of California 10 and 56.5 percent from all customers, came from companies that used syndication. GX 12 1059. Bazaarvoice highlighted the benefits of increased syndication as a reason for the 13 acquisition. Syndication is and will increasingly be a barrier to entry for R&R platforms. 14 While there are reasons companies might choose not to syndicate, syndication would be 15 extremely valuable to compete against the sophisticated product offered by Bazaarvoice. 16 Besides PowerReviews, no credible syndication competitor existed. 17 253. The Court also finds that switching costs are a barrier to entry. Bazaarvoice claims 18 that re-platforming presents an opportunity for eCommerce sites to switch R&R providers 19 with minimal incremental switching costs. Trial Tr. 1534:25-1535:18 (Goldberg). Mr. 20 Goldberg reported that close to 50 percent of large eCommerce sites surveyed said they 21 would re-platform in the next 24 months. DX1735 at 38, 44-45. But there was substantial 22 evidence that many retailers are reluctant to invest the time and money required to switch 23 R&R platforms in the absence of a demonstrable, significant benefit. Trial Tr. 1191:21- 24 1195:1 (Friedland/Build.com); Trial Tr. 1218:1-6 (Maki/Golfsmith); GX124 (Chico’s Dep: 25 27:7-16); GX168 (Onlineshoes.com Dep. 33:20-34:3); see also GX193 (Walgreens Dep. 26 61:6-62:21); 27 GX76 (Waltzinger/BBBeyond Dep. 35:6-8, 35:11, 35:16-36:8, 81:20-22, 82:4- 28 98 1 21); Trial Tr. 567:10-24, 570:12-571:1 (Waltzinger/BBBeyond); see also Trial Tr. 1012:1- 2 15 (Shapiro), GX131 (Dillard’s Dep. 65:14-24). 3 254. Svatek testified that there are costs associated with switching R&R platforms, such 4 as training, labor, and implementation. GX93 (Svatek Dep. 188:23-189:21, 189:24-191:9). 5 Even for Bazaarvoice it was “really hard” to get customers to switch from PowerReviews 6 to Bazaarvoice’s R&R platform due to the technical costs of switching and inertia. GX92 7 (Osborne Dep. 316:25-317:10, 317:13-20, 317:23-318:11); Trial Tr. 710:10-17 (Osborne); 8 GX223 at -634; Trial Tr. 210:12-211:13 (Hurt); Trial Tr. 504:14-505:6 (Defossé). 9 255. In addition, gaining the intellectual property/know how to develop and operate R&R platforms and features can be difficult and resource-intensive. Bazaarvoice 11 United States District Court Northern District of California 10 recognized that it “[w]ould take a competitor years to build a comparable solution, but then 12 we already have the customers and the network.” GX410 at -483. 13 256. Beyond the difficulty of creating a syndication network and switching costs, 14 Bazaarvoice represented to potential investors that its “moderation and analytics 15 capabilities are difficult to duplicate . . . .” GX928 at -928; Trial Tr. 888:9-889:4 (Collins). 16 The chairman of Bazaarvoice's board said when asked about Bazaarvoice’s moderation 17 capabilities, “. . . that’s important. That is a barrier.” GX491 (Meredith Dep. 197:23- 18 198:21); see also Trial Tr. 1423:11-1424:12 (Meredith). 19 257. Bazaarvoice’s moderation capabilities were “unique.” GX93 (Svatek Dep. 107:22- 20 24). Bazaarvoice attributed its sophisticated moderation capabilities to investments 21 Bazaarvoice had made in “tools, meaning IP that [Bazaarvoice] created, as well as 22 processes and, frankly, human training that were very sophisticated, very customizable and 23 configurable . . . .” GX93 (Svatek Dep. 107:22-108:13). This allowed Bazaarvoice to 24 meet “very demanding standards for large retailers and very large brands who all had very 25 sophisticated needs.” GX93 (Svatek Dep. 107:22-108:13). It would be costly for a 26 competitor to replicate these moderation capabilities. Cf. GX89 (Hurt Dep. 204:8-20); 27 Trial Tr. 106:18-107:24 (Hurt). That said, each element of the moderation process can 28 easily be performed by a new entrant, eCommerce site or outsourced to a third-party 99 1 provider of moderation services. Moderation of reviews, including the human moderation, 2 can be out-sourced to moderation vendors such as eModeration. GX64 (Tarkowski Dep. 3 27:21-28:15); GX114 (Friedland Dep. 17:6-13); DX1432 at 43; GX70 (Ahmed Dep. 4 102:15-18). There has been a proliferation of UGC (including comments sections, forums 5 and Q&A and reviews) which has created a demand for outsourced moderation services 6 and lowered the costs of such services. DX1432 at 43; Trial Tr. 106:18-21 (Hurt). 7 Moderation through a technical filter or by rules-based filtering can easily be integrated 8 into a R&R solution. GX114 (Friedland Dep. 34:6-35:2). Some eCommerce platforms, 9 such as Magento, have begun offering an integrated moderation tool with its platform. 10 United States District Court Northern District of California 11 Trial Tr. 1168:16-22 (Friedland). 258. At trial, Bazaarvoice called its analytics a “competitive strength” as Bazaarvoice 12 had “developed a – a user experience around [its] analytics capability [that] was extremely 13 modern, very easy to use” and “would allow a business user to quickly get at the heart of 14 whatever their analytics task was.” GX93 (Svatek Dep. 105:3-19, 105:24-106:8). 15 Although sites that utilize R&R often use analytics to determine how users interact with 16 R&R and other parts of their site, those analytics can be obtained from many sources. 17 DX1432 at 50, 54, 58; GX64 (Tarkowski Dep. 30:4-32:4); GX71 (Sell Dep. 12:13-15). 18 Customer testimony reflects that analytics are not a “must have” feature and that 19 preferences for analytics vary by customer. 20 259. Reputation can be a basis for customers to evaluate whether a R&R platform 21 provider is credible. GX146 (Home Depot Dep. 67:3-68:4, 69:4-15); GX72* (Lithium 22 Dep. 40:20-42:2). Bazaarvoice executives have testified that “the more clients you have, 23 the more seriously you’re taken as a business.” GX82 (Barton Dep. 51:18-52:11); see also 24 GX88 (Hossain Dep. 49:23-50:16, 51:1-2, 51:5-12, 51:14-16, 51:19-23).When searching 25 for a R&R platform provider, many customers look to the information maintained by the 26 IR500 to determine which R&R platform providers are being used by companies on the 27 IR500 list. GX130 (Dick’s Sporting Goods Dep. 30:17-32:7, 47:1-13, 47:22-48:1); GX78 (Cunningham/BJ’s Wholesale 28 100 1 Dep. 19:24-20:16). Customers generally want a R&R provider with a proven track record 2 in providing R&R and “when the price of the product is small compared to the risk it if [] 3 things go badly” customers are reluctant to try a new provider. Trial Tr. 1012:16-1013:14 4 (Shapiro). PowerReviews recognized that the combined reputation of Bazaarvoice and 5 PowerReviews serves as a barrier to entry by creating “[u]nique trust, credibility and 6 scope.” GX254 at -340. 7 260. Bazaarvoice contends that reputation is an outcome of competition, not a barrier to entry, and that firms have multiple avenues to overcome any deficit in reputation. Trial Tr. 9 1775:5-1778:21. But reputation can also be a barrier to entry for antitrust purposes. See, 10 e.g., W. Parcel Exp. v. United Parcel Serv. of Am., Inc., 190 F.3d 974, 975 (9th Cir. 1999) 11 United States District Court Northern District of California 8 (“entrenched buyer preference” may be a barrier to entry); Image Technical Services, Inc. 12 v. Eastman Kodak Co., 125 F.3d 1195, 1208 (9th Cir. 1997) (same); contra Am. Prof’l 13 Testing Serv., Inc. v. Harcourt Brace Jovanovich Legal & Prof’l Publications, Inc., 108 14 F.3d 1147, 1154 (9th Cir. 1997) (“reputation alone does not constitute a sufficient entry 15 barrier in this Circuit”) (emphasis added). 16 261. Bazaarvoice noted that Pluck, Lithium and Gigya have established relationships 17 with major clients for the provision of other social commerce tools, that they have been 18 able to strike relationships with major eCommerce platforms, and that these partnerships 19 have further bolstered the reputations of these competitors. But their reputation is not in 20 the R&R market, where they are not doing well competitively. 21 262. Witnesses’ pre-acquisition statements about “barriers to entry” contradicted 22 Bazaarvoice’s position at trial, so they tried to explain them away. For example, in the 23 documents prepared for investors and for Bazaarvoice’s road show, Bazaarvoice listed 24 several “barriers to entry.” GX0966 at 64-66; Trial Tr. 257:12-259:3. But Hurt at trial 25 said he used the term “barriers to entry” as a way of illustrating Bazaarvoice’s value 26 proposition. Trial Tr. 224:4-21, 257:12-259:3, 236:18-237:19 (Hurt). This is why 27 “culture” is identified as a “barrier to entry” along with “syndication,” he said. Trial Tr. 28 224:4-21, 135:1-19 (Hurt); GX0966 at 64-66. 101 1 263. Similarly, Luedtke prepared a presentation in which he described some 2 “Combination Benefits” including “Unique trust, credibility, and scope in emerging social 3 software category.” GX254 at -340. Under that bullet, he listed “[b]arrier to entry for sub- 4 scale firms” as a benefit. He testified that he used “barrier to entry” colloquially to 5 describe the fact that after the acquisition, Bazaarvoice and PowerReviews would have a 6 brand name and credibility in the marketplace (as described in the “[u]nique trust, 7 credibility and scope in emerging social software category”) and would be able to upsell a 8 wide range of social commerce tools to its customer base. Trial Tr. 334:23-338:2; 338:8- 9 20 (Luedtke). 10 264. In the same vein, PowerReviews employees prepared two exhibits for Ken United States District Court Northern District of California 11 Comée’s meeting with Bazaarvoice in April 2012 that showed that PowerReviews believed 12 the combination would erect barriers to entry. GX0610 and GX0612. Comée testified that 13 neither was presented to Bazaarvoice, Comée Tr. 144:24-145:2, and that he only reviewed 14 the financial information in the slides, not the ones pertaining to the combination benefits. 15 GX84 (Comée Dep. 162:22-24); Trial Tr. 1842:10-14 (Comée). The Court finds that the 16 pre-acquisition statements described above speak for themselves. 17 265. In sum, the barriers to enter the R&R platforms market in the United States 18 presented by syndication and switching costs, coupled with Bazaarvoice’s other significant 19 competitive advantages, including its reputation and customer list, as augmented by its 20 acquisition of PowerReviews, are formidable. 21 C. It is probable that the transaction will substantially lessen competition and result in higher prices 22 i. 23 24 25 26 266. Economic Testimony Supports the Conclusion That the Transaction Will Lead to Substantially Higher Prices for Many Bazaarvoice and PowerReviews Customers The merger of Bazaarvoice and PowerReviews is likely to result in significant anticompetitive unilateral effects. A “unilateral” effect is one that arises solely through the 27 28 102 1 altered behavior of Bazaarvoice without the necessity of coordinated behavior with other 2 R&R platform suppliers. Trial Tr. 911:6-25 (Shapiro); GX983* at 56-57. 3 267. Economic theory predicts that the merger will result in significant unilateral effects 4 for customers that viewed Bazaarvoice and PowerReviews as the most attractive suppliers 5 of R&R platforms and for whom the third-best supplier has a distinctly inferior product 6 compared to the weaker of the two merged parties. Trial Tr. 916:12-918:1 (Shapiro); 7 GX983* at 57-59. This theoretical prediction was supported by evidence at trial, including 8 in the win/loss data in Bazaarvoice’s Salesforce database and in the data compiled from 9 “how the deal was done” emails created by Bazaarvoice sales people (HTDWD). While both data sets may be incomplete, as Bazaarvoice charges, there was no policy within 11 United States District Court Northern District of California 10 Bazaarvoice of only recording certain competitors in the data or taking other action which 12 would create a bias in the data. See, e.g., Trial Tr. 204:5-8 (Hurt); Trial Tr. 555:14-25 13 (Defossé). The Court finds that it is reliable. 14 268. The win/loss data in Bazaarvoice’s Salesforce database tracks instances in which 15 Bazaarvoice won or lost business. GX92 (Osborne Dep. 57:13-59:21); see also Trial Tr. 16 732:18-733:19 (Osborne). Although Osborne testified that the Salesforce database was not 17 “dependable,” he admitted that the data contained in the database was accurate enough to 18 update the management and the Board of Directors about the sales pipeline. Trial Tr. 19 729:19-23, 733:20-734:5 (Osborne). 20 269. Dr. Shapiro examined 480 R&R platform sales opportunities in North America 21 identified in the Bazaarvoice Salesforce database that were created between August 2005 22 and July 2012 and that contained information about competitors. GX983* at 61 and Ex. 23 19; Trial Tr. 984:5-20 (Shapiro); GX1044. Dr. Shapiro concluded that PowerReviews was 24 identified as a competitive alternative in 75 percent of the opportunities. GX983* at 61 25 and Ex. 19; Trial Tr. 985:3-7 (Shapiro); GX1044. The next closest alternative, in-house 26 solutions, appears in 18 percent of the opportunities. GX983* at 61 and Ex. 19; Trial Tr. 27 985:8-10 (Shapiro); GX1044. No other alternative appears in more than three percent of 28 103 1 the opportunities. GX983* at 61 and Ex. 19; see also Trial Tr. 985:11-15 (Shapiro); 2 GX1044. 3 270. The results are similar when the analysis is limited to customers in the IR500. As 4 shown by the chart below, Dr. Shapiro concludes that PowerReviews was identified as a 5 competitive alternative solution in 83 percent of the opportunities. The next closest 6 alternative, in-house solutions, appears in 15 percent of the opportunities. No other 7 Frequency of Competitors in Bazaarvoice Win/Loss Opportunities 8 IR500, Core R&R Products Only 9 Competitor PowerReviews Internal Build Pluck Viewpoints Reevoo Lithium Expo Jive Zuberance "Other" Total Opportunities GX1045 10 United States District Court Northern District of California 11 12 13 14 Opps 121 22 4 4 1 0 0 0 0 2 146 Frequency 83% 15% 3% 3% 1% 0% 0% 0% 0% 1% 15 16 17 18 19 20 21 22 23 24 25 26 27 competitive alternative appears in more than three percent of the opportunities. GX983* at 61 and Ex. 20; GX1045*. 271. The Bazaarvoice HTDWD email dataset produces similar results. Bazaarvoice salespeople were generally expected to send a How the Deal Was Done (“HTDWD”) email to a specific group email list at the company at the close of each transaction explaining how the transaction was accomplished. It is a useful data set to help evaluate the competition faced by Bazaarvoice. 272. Although Bazaarvoice had no formal process to ensure the HTDWD emails were complete or accurate, Bazaarvoice did not instruct its employees to skew the reports to overstate engagements involving PowerReviews. Trial Tr. 728:19-729:5, 734:6-16 (Osborne). There was no bias and the fact that it, like most data, is incomplete does not render the analysis unreliable, especially when it is consistent with the other evidence. 28 104 1 Trial Tr. 2059:7-2061:4 (Shapiro); cf. Trial Tr. 204:5-8 (Hurt); Trial Tr. 555:14-25 2 (Defossé). 3 273. As shown in the following chart, Dr. Shapiro examined 143 Bazaarvoice HTDWD 4 emails sent between September 2008 and July 2012 in which a competitor was identified 5 and concluded that PowerReviews accounted for 80 percent of the references to 6 competitors in these emails. The next closest alternative identified was in-house solutions, 7 which accounted for 12 percent of the competitive references. No other competitive 8 alternative appears in more than three percent of the emails. Trial Tr. 986:12-987:1; 9 GX983* at 62, Ex. 26 and Ex. 27; GX1047; GX1048*. 10 Competitor Counts in Bazaarvoice "How the Deal was Done" Documents United States District Court Northern District of California 11 Ratings and Reviews Deals Only 12 Competitor PowerReviews In-house Pluck Shopzilla Viewpoints Buzz Metrics Gigya Reevoo Trip Advisor 13 14 15 16 17 Opps 114 17 4 2 2 1 1 1 1 Total Competitor Counts GX1048 18 Frequency 80% 12% 3% 1% 1% 1% 1% 1% 1% 143 19 20 21 22 23 24 25 26 27 28 274. Dr. Shapiro identified three types of customers “that are most likely to be vulnerable and harmed” by the merger. Trial Tr. 998:18-999:1 (Shapiro). These three groups are Bazaarvoice customers, PowerReviews legacy customers, and customers who place a high value on syndication. Trial Tr. 998:18-999:1 (Shapiro). 275. Dr. Shapiro concludes that the unilateral price effects of the Bazaarvoice/PowerReviews merger will be most pronounced for their existing customers when those customers’ current contracts come up for renewal. Trial Tr. 999:6-24 (Shapiro); GX983* at 63. These customers have already revealed their preference for a 105 1 commercial R&R platform and because of the merger they have lost the ability to play 2 Bazaarvoice and PowerReviews off one another to get a lower price. Trial Tr. 999:15-23 3 (Shapiro); GX983* at 63. 4 276. One mechanism Bazaarvoice could employ to unilaterally raise prices would be to 5 eliminate the PowerReviews product from a customer’s choice set and migrate legacy 6 customers from the PowerReviews enterprise platform to the higher-priced Bazaarvoice 7 product. Trial Tr. 1003:14-1004:11 (Shapiro); GX983* at 64. While the customer may get 8 additional features, these would not be features that the customer would otherwise choose 9 if PowerReviews’s lower-priced alternative were still an option. Trial Tr. 1003:20-1004:4 10 United States District Court Northern District of California 11 (Shapiro); GX983* at 65 and fn.307; see also GX293 at -118. 277. Bazaarvoice planned to migrate the largest PowerReviews customers to the higher- 12 priced Bazaarvoice platform. In proposing an acquisition of PowerReviews in April 2011, 13 Bazaarvoice’s co-founder Brant Barton described this as one “Pro” in favor of the deal: 14 “[w]e could migrate their Tier 1 customers to our platform . . . .” GX514 at -810. In 15 addition, a May 2012 Bazaarvoice presentation points out that average revenue from IR500 16 customers at PowerReviews in terms of annualized cumulative billings is 83% below 17 Bazaarvoice’s, adding: 18 [The merger p]rovides us with the opportunity to increase revenue from [PowerReviews’] existing clients via migration to the Bazaarvoice platform a [sic] higher price points in return for greater features and functionality that [sic] those of [PowerReviews]. . . . 19 20 GX332 at -291; Trial Tr. 775:23-776:11 (Collins). 21 22 278. As a result of the acquisition, Bazaarvoice gained the ability to unilaterally charge 23 higher prices for its new customers who do not consider in-house (or fringe competitor 24 solutions) to be a viable option, as compared to its pre-merger new customers. GX983* at 25 63; see also Trial Tr. 2050:6-2051:19 (Shapiro). In many instances, Bazaarvoice can 26 effectively price discriminate against such customers because it collects detailed 27 information about a customer’s requirements and budget constraints during the sales 28 process. See Trial Tr. 998:2-17 (Shapiro). The more effectively Bazaarvoice can engage 106 1 in targeted “price discrimination” based on customer attributes, the greater the unilateral 2 competitive effects. GX983* at 63; see also Trial Tr. 997:18-998:1 (Shapiro). 3 279. Customers for whom syndication is particularly important, especially brand 4 customers, are also likely to be harmed by the merger. Trial Tr. 998:18-999:1, 1001:5- 5 1002:13 (Shapiro). For brand customers looking for a syndication network with significant 6 retail customers, PowerReviews was the closest and only credible alternative to 7 Bazaarvoice. Trial Tr. 1001:5-1002:13, 2051:20-2052:13 (Shapiro). After the merger, 8 these customers have no alternative R&R platform with a significant syndication network 9 and therefore these customers have lost significant bargaining leverage. Trial Tr. 1001:51002:13, 2051:20-2052:13 (Shapiro). This applies both to new customers as well as legacy 11 United States District Court Northern District of California 10 customers who now have access to a larger syndication network, as “Bazaarvoice will seek 12 to capture that value in the price they charge. The customers’ return will depend on what 13 their bargaining leverage is, and that’s been weakened.” Trial Tr. 1001:5-1002:13 14 (Shapiro); see also Trial Tr. 2051:20-2052:13 (Shapiro). 15 280. Other commercial suppliers of R&R platforms are not sufficiently close substitutes 16 to Bazaarvoice’s platform to prevent a significant post-merger price increase. See, e.g., 17 GX1181 at -158 (PowerReviews is “the only credible player in the space”). In April 2011, 18 Barton discussed the absence of competitive alternatives for customers, concluding that 19 Bazaarvoice would “retain an extremely high percentage of [PowerReviews] customers,” 20 because available alternatives for disgruntled customers were “scarce” and “low-quality.” 21 GX514 at -810; GX82 (Barton Dep. 208:20-212:14); Trial Tr. 647:19-649:15 (Barton). 22 After the acquisition, a customer made the same point to Luedtke: “from a retailers stand 23 point I have to say I am a little concerned that there is now only one option for customer 24 reviews.” GX892 at -323. Similarly, BBBeyond came to the conclusion that there were 25 only two options for them in 2009, Bazaarvoice and PowerReviews, and came to the same 26 conclusion in 2012. GX76 (Waltzinger/BBBeyond Dep. 59:21-60:3, 60:6-11, 96:7-17); 27 Trial Tr. 571:18-25 (Waltzinger/BBBeyond); see also GX98 (American Eagle Dep. 21:10- 28 14, 21:17-19). 107 ii. 1 2 281. Post-acquisition evidence regarding price To establish a violation of Section 7 of the Clayton Act, the government need not 3 prove that the merger has resulted in higher prices or had other anticompetitive effects. 4 Rather, the government must show a “reasonable likelihood” of anticompetitive effect in 5 the relevant market. See e.g., Penn-Olin Chem. Co., 378 U.S. at 171 (a Section 7 violation 6 is shown when “the ‘reasonable likelihood’ of a substantial lessening of competition in the 7 relevant market is shown”) (citing E. I. du Pont de Nemours & Co., 353 U.S. at 592); 8 Oracle Corp., 331 F. Supp. 2d at 1109-10 (“Section 7 does not require proof that a merger 9 or other acquisition [will] cause higher prices in the affected market. All that is necessary is that the merger create an appreciable danger of such consequences in the future.”). 11 United States District Court Northern District of California 10 Because the acquisition occurred more than fifteen months prior to trial, both parties 12 offered evidence of the effect of the merger on pricing. The Court is obligated to consider 13 the weight this evidence should be given. 14 282. The evidence regarding price increases or decreases post-acquisition, discussed 15 below, is inconclusive and the Court does not give it much weight. In any event, the 16 probative value of post-acquisition evidence is particularly limited “whenever such 17 evidence could arguably be subject to manipulation.” Chicago Bridge, 534 F.3d at 435 18 (emphasis in original). In this regard, the Department of Justice informed Bazaarvoice that 19 it had opened an investigation just two days after Bazaarvoice acquired PowerReviews. 20 Several internal communications at Bazaarvoice after the acquisition made clear that 21 Bazaarvoice was aware of the Department of Justice’s interest. See, e.g.,GX495 at -301; 22 GX527; GX1225 at -047; GX352* and Trial Tr. 781:8-782:3 (Collins). In response to a 23 Huffington Post article on the acquisition in June 2012, prognosticating that Bazaarvoice 24 has a monopoly and that prices will rise, Bazaarvoice’s Director of Communications wrote 25 that “[w]hatever we come up with will need to be vetted by legal so we avoid any anti-trust 26 gotchas.” GX 495 at -301. The post-acquisition evidence regarding pricing and the effect 27 of the merger is reasonably viewed as manipulatable and is entitled to little weight. 28 108 1 283. Bazaarvoice has raised prices for some customers post-acquisition. It argues that 2 post-merger price increases occurred due to the addition of additional features or improved 3 functionality. It has also lowered prices or kept them at the same level for other customers. 4 The government argues that these customers saw decreases in the features provided. No 5 pronounced trend was clear to the Court. 6 284. The government pointed to several examples of customers receiving price 7 increases. , a PowerReviews customer, complained to Bazaarvoice after 8 the transaction: “[s]o we’re paying more (at threat of cutting us off during our discussion, 9 which was a total turn-off after years of a great partnership) and getting less. We need to get back to the great service that PowerReviews provided.” GX338* at -342. When 11 United States District Court Northern District of California 10 Bazaarvoice attempted to explain the price increase as not related to the merger, 12 13 14 15 16 17 18 said that: we were paying for four sites and were then told one more site would be another was really treated like dirt and the launch of our was held hostage by a longterm partner, or should I say by the company that bought our longterm partner . . . . None of this sort of thing happened when I was with PowerReviews, and this all reinforced my belief that I made the right decision X years ago when I selected PR over BV.GX338* at 340. But Dr. Shehadeh countered that the increase in price that was offered to 19 was just what PowerReviews would have done to increase prices at renewal. Trial Tr. 20 1894:2-1895:11 (Shehadeh). When Collins became aware of 21 he pushed back against the price increase because Bazaarvoice’s “long-term value 22 ’ concerns, proposition” was different than PowerReviews’s and rested on the ability to make money 23 on the data generated from each customers. Trial Tr. 1894:2-1895:11 (Shehadeh); 24 25 DX1892; DX1894. Collins testified that was mainly concerned about 26 Bazaarvoice’s higher moderation standards and the fact that 27 to remove authentic negative reviews as they had done when working with PowerReviews. 28 Trial Tr. 797:6-10 (Collins); DX1892 at -520. 109 might be able 1 285. 2 , a PowerReviews customer, said that after the merger: [o]ur first experience as a BV customer was to receive a increase for our renewal in July – not a very good start considering we weren’t receiving any meaningful upgrades to our service. My team felt that we had been put over a barrel since we believed that the two premier players in this space had been reduced to one and to make matters worse BV was taking advantage of this situation with heavy handed pricing.” 3 4 5 6 7 GX337* at -523-524. added that “your company’s actions relative to pricing . . 8 . raised a huge red flag to the DOJ that competition was being diminished.” Id. at -524 9 (emphasis in original). 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 286. eventually complained to Luedtke and got the increase reduced to eight percent with no meaningful improvements to service. Zach Bolian, a financial analyst at Bazaarvoice acknowledged to Gary Skinner, another Bazaarvoice financial analyst, that “there was a price increase with no incremental value . . . .” GX337* at -523. Skinner followed up with a directive to “stop this immediately” and that he “will look for others that have been increased since the transaction closed.” GX337* at -523. 287. Staples reached out to Bazaarvoice expressing its concerns with the transaction after hearing that in Staples’ renewal negotiations Bazaarvoice attempted to impose a “2x increase in the core cost of PR service to staples.com with zero added value provided to us.” GX257; GX692*. But Bazaarvoice noted that Staples received several additional features including an upgrade of all its sites to the ESS platform (which includes Loyalty, Q&A, Crowdsource, Q&A Ask Product Owner and Premium Support). It also considerably expanded services for its foreign sites. DX1851; Trial Tr. 437:16-439:16 (Luedtke). 288. Dillard’s received a price increase of approximately without any corresponding increase in service or functionality. GX131 (Dillard’s Dep. 44:5-45:6, 45:925, 46:4-6, 46:13-47:15, 52:15-53:1, 72:23-73:18). 27 28 110 1 289. Just prior to the merger, was paying Bazaarvoice per year for its 2 R&R platform, limited to 10,000 SKU’s, and was in the process of renegotiating its 3 contract. PowerReviews offered 4 of 5 could be tricky for us to sell an uptick in fees.” GX1149* at -586; see also GX120. A 6 salesperson at Bazaarvoice then asked, “Can someone communicate back to the PR team 7 that ‘we’ve got this one?’” GX1149* at -586. 8 Bazaarvoice, agreeing to a significant price increase. GX821* at -887. 9 290. its R&R platform for 100,000 SKU’s at a price . After the merger, Bazaarvoice learned of PowerReviews’s offer, and noted, “it ultimately renewed with Bed Bath & Beyond (“BBBeyond”) received a post-merger price increase. In 2009, BBBeyond decided to add R&R to its two online retail websites because it was at a 11 United States District Court Northern District of California 10 competitive disadvantage without R&R and believed R&R would improve conversion. 12 GX76 (Waltzinger/BBBeyond Dep. 20:17-24, 21:8-15, 21:18); Trial Tr. 561:25-562:12 13 (Waltzinger/BBBeyond). BBBeyond determined that it would not make sense to build its 14 own in-house solution and that there were only two R&R platform providers that met its 15 needs: Bazaarvoice and PowerReviews. GX76 (Waltzinger/BBBeyond Dep. 21:19-22:6, 16 22:10-20, 22:23-23:19, 23:22-24:7, 24:11-16, 25:16-18, 90:12-14, 90:17-91:3); Trial Tr. 17 562:13-564:3, 564:5-8, 564:21-23, 571:2-12 (Waltzinger/BBBeyond). In its negotiations 18 with Bazaarvoice, BBBeyond was able to obtain a free trial year, which represented a cost 19 savings of 20 GX1005; GX548. 21 291. by threatening to pursue a deal with PowerReviews. GX659 at -477; BBBeyond renewed its contract with Bazaarvoice after the merger. Bazaarvoice 22 attempted to impose a /year price increase and BBBeyond eventually agreed to a 23 smaller increase of 11 percent. GX1026 at -847; GX76 (Waltzinger/BBBeyond Dep. 32:3- 24 5, 32:8-16, 32:19-21, 32:25-33:23, 34:8-11, 34:14-35:5); Trial Tr. 565:18-567:9 25 (Waltzinger/BBBeyond). BBBeyond believes that the presence of a credible competitor 26 “helps out the marketplace” during negotiations and today does not believe there is another 27 alternative to Bazaarvoice that would meet its R&R platform needs. GX76 28 (Waltzinger/BBBeyond Dep. 98:2-24, 99:4-13, 106:23-107:23, 108:5-17, 109:13-17, 111 1 109:21-110:4, 110:12-15, 110:20-111:10); Trial Tr. 572:1-574:18, 581:9-13 2 (Waltzinger/BBBeyond). 3 292. But Bazaarvoice pointed out that BBBeyond had not canvassed to see whether 4 there are additional R&R providers beyond PowerReviews or Bazaarvoice, Waltzinger Tr. 5 98:2-5, and its deponent admitted that he does not keep abreast of developments in the 6 R&R sphere. GX76 (Waltzinger Dep. 100:15-21). In 2012, BBBeyond did not investigate 7 any other R&R vendors besides Bazaarvoice and had never performed a specific 8 estimation of the costs of switching R&R platforms. Id. at 81:20-82:24. BBBeyond had 9 never estimated the initial or ongoing costs for building a R&R solution in-house. 10 293. In addition, according to Bazaarvoice, BBBeyond received additional services from United States District Court Northern District of California 11 Bazaarvoice in 2012. BBBeyond’s historic contract was structured so that there were 12 volume “tiers” of costs for moderating a certain number of reviews. As part of the 13 negotiations in 2012, Bazaarvoice stated that it based pricing on a projection of the number 14 of reviews BBBeyond would accumulate. Bazaarvoice proposed a price increase of 33 15 percent which reflected the historic and future performance while eliminating tiers and 16 allowing BBBeyond unlimited review volume, but BBBeyond was able to negotiate that 17 increase down to 11 percent without seeking competitive bids from other vendors. Trial 18 Tr. 1131:17-1133:9 (Godfrey). In addition, BBBeyond received additional services 19 launched R&R in Canada at no additional costs. GX76 (Waltzinger Dep. 71:7-19). 20 294. Bazaarvoice argued that its sales data indicated that the majority of its R&R 21 customers did not experience an increase in prices from June 2012 to March 2013, even if 22 their contracts came up for renewal post-merger. DX1433 at ¶ 239. It says that 14 percent 23 of its customers experienced an increase in price between the merger and trial. DX 1433 at 24 ¶ 239, Exhibits 4, 5. 25 26 27 28 295. Bazaarvoice pointed to among others, which received substantially better pricing post-merger. Pre-merger, received R&R and Facebook applications for total ASF. DX1433 at ¶ 93. renewed their contract in August 2012 for , with no incremental ASF. DX1433 at ¶ 93. 112 1 296. The parties quarreled whether price reductions were also accompanied by service 2 reductions. For example, Vitamin Shoppe’s PowerReviews price decreased in its most 3 recent contract. But the government showed that the lower price came with significant 4 service reductions, including a downgrade from a standard to basic service package, 5 removal of Facebook Discovery and Community Services, and reductions in moderation 6 and technical response service levels. Compare DX1796 with GX1221. Similarly, while 7 GameStop was initially cited as an example of a post-merger price reduction because of 8 competition, the evidence showed that the price decrease was actually associated with a 9 reduction in service. Trial Tr. 1140:1-8, 1141:9-1142:21 (Godfrey); GX1231 at 1. 10 297. Bazaarvoice points to its post-merger financial performance to downplay the United States District Court Northern District of California 11 advantage it gained since acquiring PowerReviews. It says that competition has intensified 12 and “constant pressure” on pricing remains the norm. Trial Tr. 892:13-892:21 (Collins). It 13 argues that since the merger, Bazaarvoice has “had to accept significant [demands for price 14 reductions from customers for reasons] more significant and more powerful than what 15 PowerReviews ever could have done, and that those [market] forces have accelerated and 16 are continuing to impact [Bazaarvoice] in material ways.” Trial Tr. 879:1-879:19 17 (Collins). 18 298. To prove this, Bazaarvoice cites to its lost revenues, or “churn” data. It claims that 19 the amount of “churn” has doubled in the year since the merger as compared to the year 20 before the merger ($10.8 million v. $21 million). DX1870; DX1085 at -252; Trial Tr. 21 855:5-856:1, 879:20-880:4. Its initial analysis immediately following the merger showed 22 that competitive losses had dropped, but this seemed to have changed by the trial. Trial 23 Tr. 876:3-878:25 (Collins); GX1216 at 17-18. Other negative measures included an 24 increase in the number of mid-contract adjustments, DX1092; DX1843 at -809, and a 25 decrease in its retention rate at renewal in the second and third quarters of fiscal year 2013. 26 In the previous two years, the renewal retention for brands was 91 percent; in Q2 and Q3 27 fiscal year 2013, it was 80 percent. Similarly, the renewal retention rate in retail dropped 28 from a two year average of 88 percent to 79 percent in Q2 and Q3 of fiscal year 2013. 113 1 DX1081 at -902. Bazaarvoice says that its growth rate is roughly half of what it was in the 2 preceding years and that it did not meet its internal sales objectives for fiscal year 2013. 3 Trial Tr. 854:3-854:5 (Collins). 4 299. There was insufficient evidence to identify any specific factor to account for these 5 results. This analysis did not include the impact of new clients. It counted a customer’s 6 switch from PowerReviews to Bazaarvoice as churn, as well as migration from 7 PowerReviewsExpress to PowerReviews or Bazaarvoice. Churn also is affected by events 8 that happen to customers over which Bazaarvoice could have no control, like bankruptcies. 9 300. And while Bazaarvoice is not profitable, it has never been profitable. Trial Tr. 189:2-11, 220:21-24, (Hurt), 854:16-855:4 (Collins). That has not stopped investors from 11 United States District Court Northern District of California 10 supporting its successful IPO and follow-on funding. It is not surprising that the merged 12 company has more churn than before, given the difficulties of melding new customers, 13 different products and new employees into the existing organization. Of course, it is also 14 typical for a young technology company to prioritize a growth strategy over immediate 15 profitability and to have bumps in its road to success. The Court does not conclude that 16 the financial reports since the merger imply that the merger will not have anticompetitive 17 effect. 18 301. Bazaarvoice attacks Dr. Shapiro’s ultimate conclusion that it is probable that the 19 transaction will substantially lessen competition in several ways. First, it says that Dr. 20 Shapiro concedes price effects will vary from customer to customer, Trial Tr. 997:18-998:1 21 (Shapiro), but his opinion does not depend on “customer-by-customer” price 22 discrimination. Trial Tr. 1038:10-19 (Shapiro). Bazaarvoice argues that such 23 discrimination would be required to make a hypothetical price increase profitable because 24 many similarly situated customers testified that they have competitive alternatives other 25 than Bazaarvoice. 26 302. Further, Bazaarvoice argues that Dr. Shapiro performed no analysis to show harm 27 to in-house customers from the merger. Trial Tr. 1879:5-12 (Shehadeh), 2071:14-20 28 (Shapiro). Dr. Shapiro acknowledges that customers can threaten to turn to in-house 114 1 solutions and other competitors, Trial Tr. 999:15-20, 1000:19-25 (Shapiro), and concedes 2 that in-house development is “a significant choice” for many customers. Trial Tr. 3 1022:22-23 (Shapiro). He did not identify any customer who threatened to go in-house 4 and was not taken seriously. Trial Tr. 1074:13-18 (Shapiro). Dr. Shapiro did not identify 5 any customer who cannot use Pluck, Gigya, and in-house as leverage in negotiating with 6 Bazaarvoice. Trial Tr. 2073:18-2074:8 (Shapiro). 7 303. Bazaarvoice also contends that Dr. Shapiro’s claim that legacy Bazaarvoice customers will be harmed is incorrect because PowerReviews was a weak competitor. And 9 Dr. Shapiro’s claim that legacy PowerReviews customers will be harmed is incorrect, it 10 says, because (i) such customers could use in-house development and other competitors 11 United States District Court Northern District of California 8 for better leverage against Bazaarvoice than PowerReviews would have provided going 12 forward, (ii) Bazaarvoice is offering small-to-medium sized customers with services 13 comparable to PowerReviews’s Enterprise platform, and (iii) Bazaarvoice has changed its 14 pricing model to charge lower prices to the types of small-to-medium sized customers 15 PowerReviews served. Trial Tr. 1889:6-1891:25 (Shehadeh). In addition, it says that Dr. 16 Shapiro’s claim that customers who value syndication will be harmed is incorrect because 17 PowerReviews did not have a significant syndication presence and syndication can be done 18 separately from R&R. Trial Tr. 1896:5-1900:3 (Shehadeh). 19 304. Ultimately, these arguments are not persuasive. By any measurement, 20 PowerReviews had at least five times the market share of all other commercial competitors 21 combined (excluding in-house solutions), and it has not come close to being replaced. Its 22 customer base represented the only meaningful source of content from which Bazaarvoice 23 could secure an insurmountable advantage for the syndication network effect. For many 24 retailers and manufacturers, in-house R&R platforms are not sufficiently close substitutes 25 to Bazaarvoice’s or PowerReviews’s platforms to impede a post-merger price increase by 26 Bazaarvoice. GX365* at -478; GX939 at -551; GX678; Trial Tr. 1188:12-1189:12 27 (Friedland/Build.com). While existing competitors and the potential of going in-house 28 may well exercise some price restraint, Bazaarvoice’s enhanced market power, the 115 1 significant barriers to entry, and the weakness of its remaining competitors leads the Court 2 to agree with Dr. Shapiro’s central conclusion that it is more likely than not that the merger 3 will have anticompetitive effect. 4 iii. Customer Testimony 5 6 305. Bazaarvoice relied heavily on the fact that none of the more than100 current, 7 former and potential customers who testified in this case believed that the acquisition had 8 harmed or would harm them. The Court finds that the customers were the most credible 9 sources of information on their need for, use of and substitutability of social commerce products, as well as regarding their companies’ past responses to price increases. But 11 United States District Court Northern District of California 10 customers generally do not engage in a specific analysis of the effects of a merger. Trial 12 Tr. 2077:15-2078:3 (Shapiro); cf. GX103 (B&H Foto Dep. 42:20-25). Many of them had 13 given no thought to the effect of the merger or had no opinion. They lacked the same 14 information about the merger presented in court, including from the economic experts. 15 Their testimony on the impact and likely effect of the merger was speculative at best and is 16 entitled to virtually no weight. 17 306. Products such as R&R platforms are relatively inexpensive in comparison to a 18 company’s operating budget and have relatively long contracts. Considering competitive 19 alternatives for R&R is not part of customers’ day-to-day activities and most do not have 20 much current information about the prospective effect of a merger. Trial Tr. 2045:3- 21 2046:10 (Shapiro). As Dick’s Sporting Goods testified in this case, “[i]f I’m not in the 22 market to buy ratings and review services at the moment, I’m not up to speed on that 23 current service, managing hundreds of different vendors. I look at them when they’re 24 needed, not continually research everything that’s out there all the time.” GX130 (Dick’s 25 Sporting Goods Dep. 32:18-21, 32:24-33:5); see also GX170 (Overstock.com Dep. 41:5- 26 10) (when asked whether Overstock had been harmed by the merger, testified “I don’t 27 really have enough information to accurately answer that”). 28 116 1 307. Many customers testified that they had never given any thought to the merger or had no opinion about it. See, e.g., Trial Tr. 1268:15-17 (Yudin/Abe’s of Maine); GX94 3 (Abercrombie Dep. 25:10-13); GX132 (Drums on Demand Dep. 25:20-26:7); GX134 4 (eHobbies Dep. 21:17-19, 38:21-39:1); GX136 (Footlocker Dep. 36:1-5, 47:7-10); GX143 5 (Guess? Dep. 44:11-14); GX154 (Johnsonville Sausage Dep. 19:20-22); GX157 (K-Swiss 6 Dep. 17:5-8); GX160 (Lord & Taylor Dep. 31:9-15, 31:18); GX164 (New Era Dep. 42:21- 7 24); GX173 (Paypal Dep. 17:5-8); GX174 (Perform Better Dep. 19:16-19); GX175 (Petco 8 Dep. 21:9-14); GX183 (Shoplet Dep. 42:23-43:1); GX141 (Golfsmith Dep. 28:16-18, 9 28:20-23, 62:12-14, 62:16-17); GX138* (Fruit of the Loom Dep. 34:12-16); GX46 (Acer 10 Dep. 67:7-8), GX105 (Belk Dep. 21:14-17), GX127 (Columbia Sportswear Dep. 30:18- 11 United States District Court Northern District of California 2 22), GX126 (Coldwater Creek Dep. 46:16-19), GX124 (Chicos Dep. 27:17-23), GX180 12 (Seagate Dep.30:10-15), GX182 (Shoebuy Dep. 58:3-8), GX186 (Signal Group Dep. 13 21:14-17), Trial Tr. 1679:14-25 (Bruns/,Southwest), GX189 (Temper-Pedic Dep. 45:6-9), 14 GX191 (UnderArmor Dep. 15:25-16:2), GX193 (Walgreens Dep. 30:25-31:6), and GX197 15 (XO Group Dep. 33:6-9). 16 308. Others were either unaware of alternatives or had conducted a limited review of 17 their alternatives. See, e.g., Trial Tr. 468:7-17 (Cunningham/BJ’s Wholesale); Trial Tr. 18 1639:11-24, 1640:5-17, 1641:3-9 (Moen/Hayneedle); GX111 (Bon-Ton Dep. 45:15-19, 19 45:22-46:9, 48:20-49:4); GX113 (The Buckle Dep. 36:10-37:1); GX115 (Cabela’s Dep. 20 21:13-17, 22:15-20); GX130 (Dick’s Sporting Goods Dep. 32:18-21, 32:24-33:10); GX136 21 (Footlocker Dep. 36:1-5); GX169 (Orchard Supply Dep. 28:10-29:1); GX183 (Shoplet 22 Dep. 42:16-22); GX79 (Toys R Us Dep. 58:4-8). 23 309. Several others did not currently use a commercial R&R platform yet gave an 24 opinion on the effects of the merger. See, e.g., GX94 (Abercrombie Dep. 18:16-20); 25 GX110 (Blue Nile Dep. 20:19-23); GX133 (eBags Dep. 49:7-13); GX147 (HSN Dep. 26 45:11-20); GX176 (Planet DJ Dep. 27:2-9); GX183 (Shoplet Dep. 22:20-22, 22:24-25, 27 38:16-19); GX188 (Systemax Dep. 31:8-12, 31:15-23); cf. GX169 (Orchard Supply Dep. 28 27:16-18, 27:21-28:3) (dropped R&R due to a shift in focus away from eCommerce, but if 117 1 they add it back in the future he would have “some concerns if these were still the only two 2 players that had gone to one in that space.”). 3 D. Bazaarvoice has not demonstrated that merger-specific efficiencies will overcome the anticompetitive effects of the merger 4 5 310. Bazaarvoice contends that the acquisition of PowerReviews by Bazaarvoice 6 resulted in additional value to Bazaarvoice’s retail and brand customers because 7 Bazaarvoice now has access to a large amount of data and will be able to provide more 8 value to its clients with additional and more powerful data analytics products, expand its 9 social commerce marketing solutions focused on the retail channel for brands, and offer additional advertising and opportunities to engage with a wider audience for brands. 11 United States District Court Northern District of California 10 DX0735; Trial Tr., 1391:6-1392:5, 1429:5-1430:19; DX0649 at BZ-01448284; Trial Tr. 12 791:9-793:3, 836:21-839:18, 840:7-841:21. While the merger absolutely benefits 13 Bazaarvoice on a number of levels, it is unclear that its customers will reap those benefits 14 absent price increases or, in some cases, that they would not have received those benefits if 15 the merger had not occurred. 16 311. Bazaarvoice justified the merger in part as an attempt to better compete in the 17 online advertising space. Trial Tr. 271:9-12 (Hurt). It has a preexisting media product. 18 Trial Tr. 683:9-11 (Barton). It also made a separate acquisition to expand in the 19 advertising space. Trial Tr. 271:13-17 (Hurt). But Hurt and Collins testified that there was 20 nothing preventing Bazaarvoice from selling advertising products to PowerReviews clients 21 absent the merger. Trial Tr. 271:18-272:1 (Hurt); Trial Tr. 882:16-19 (Collins). 22 312. Bazaarvoice makes no claim that the merger reduces the marginal costs of 23 providing its services. See GX871 (Def.’s Am. Resp. to Pl. Interrogs. 3,5,6, & 13 7:25- 24 10:26); GX494 (Godfrey 30(b)(6) Dep. 19:9-22, 47:13-17); GX1004. It acknowledges that 25 it expects only limited cost savings from the acquisition. GX494 (Godfrey 30(b)(6) Dep. 26 41:8-19). Given its value-based pricing strategy, Bazaarvoice can be expected to charge its 27 customers higher prices for any improvements in product quality rather than passing the 28 benefits through to customers. Trial Tr. 1002:4-13, 1021:20-24 (Shapiro); GX984* at 27. 118 1 313. There is insufficient evidence to support Bazaarvoice’s claim that the merger has 2 enabled it to launch “architectural improvements” in its technology. Bazaarvoice ceased 3 development of the PowerReviews platform other than ongoing maintenance following the 4 acquisition. GX81 (Collins Dep. 408:14-20). Furthermore, any improvements to the 5 Bazaarvoice platform or service could have been completed without the transaction. 6 Bazaarvoice was working on these improvements prior to the merger and expected to 7 implement them successfully. GX87 (Godfrey CID Dep. 51:14-53:9); see also GX81 8 (Collins Dep. 397:11-21); GX82 (Barton Dep. 243:15-24, 244:5-246:17); GX519; Trial Tr. 9 1153:3-1154:9 (Godfrey). Bazaarvoice did not quantify how much the merger reduced the cost or time of developing architectural improvements to its technology or the value of 11 United States District Court Northern District of California 10 such improvements to consumers. GX87 (Godfrey CID Dep. 51:14-53:9); GX494 12 (Godfrey 30(b)(6) Dep. 32:16-33:3, 33:13-34:5, 52:12-53:2). 13 314. Bazaarvoice believes it must “constantly innovate to continue to be relevant.” Trial 14 Tr. 835:6-836:9 (Collins). It is developing new services such as Bazaarvoice 15 Conversations and Bazaarvoice Connections. Trial Tr. 785:22-786:15, 788:20-790:4 16 (Collins). It has increased its spending on R&D since the acquisition of PowerReviews, 17 spending about $40 million in 2013 on R&D; in the last two and a half years Bazaarvoice 18 has gone from approximately 30 programmers to about 150 today. Trial Tr. 861:18-862:11 19 (Collins). But there is no reason to believe that the merger caused any of this spending. 20 Instead, as Bazaarvoice asserted at trial, it has a need to innovate that “comes from [its] 21 need to create value for [its] clients first and foremost, and the marketplace generally is 22 competitive and very dynamic for all things eCommerce and Internet.” Trial Tr. 759:18- 23 759:25 (Collins). 24 315. Bazaarvoice says that the larger volume of data it will control post-merger will 25 enable unspecified improvements in the company’s “analytics” tools. However, it 26 acknowledges that it could have shared data with PowerReviews absent the merger and, in 27 the future, Bazaarvoice “fully expect[s]” to share data sets with other online software 28 providers to expand analytic power. GX83 (Collins 30(b)(6) Dep. 46:15-47:8). 119 1 Bazaarvoice asserts that the acquisition resulted in the combined company “bringing 2 together the very best technologies and capabilities from both companies” with the goal of 3 developing “a next-generation platform” for its customers. DX15 at -231. But there is no 4 verifiable evidence that the combination of the Bazaarvoice and PowerReviews analytics 5 research and development efforts constitute a merger-specific benefit. Both companies 6 could have increased their efforts without the merger (see GX83 (Collins 30(b)(6) Dep. 7 40:9-21)) so it is a matter of conjecture whether the potential benefits from such combined 8 efforts will outstrip those if the companies had remained separate. Even if the acquisition 9 did lead to a merger-specific improvement in Bazaarvoice’s analytics capabilities, Bazaarvoice reasonably can be expected to charge a price commensurate with that greater 11 United States District Court Northern District of California 10 capability, eliminating any net benefits to consumers. See GX984* at 27; GX83 (Collins 12 30(b)(6) Dep. 50:9-14, 50:17-51:7); GX117 (Parsons 30(b)(6) Dep. 25:7-17). 13 316. Bazaarvoice began syndicating content from Bazaarvoice customers to legacy 14 PowerReviews customers in October 2012. Bazaarvoice says that the process of utilizing 15 content from Bazaarvoice brands into the PowerReviews platform for syndication resulted 16 in Bazaarvoice addressing deficiencies in the product data feeds that it used, which will 17 lead to “greater automated catalog matching.” DX15 at -027. The acquisition increased 18 Bazaarvoice’s reach to 1.8 billion monthly visitors, providing “new opportunities for 19 content syndication.” DX15 at -231. However, the ability to syndicate R&R content 20 between PowerReviews and Bazaarvoice customers or “cross network syndication” is not a 21 merger specific efficiency. Such syndication was feasible and was occurring for some 22 customers prior to the merger. See, e.g., GX493 (Parsons Dep. 82:20-83:13); GX117 23 (Parsons 30(b)(6) 34:9-35:4); GX4 at -460; GX699* at -311; GX276 at -765; cf. GX1009 24 (Curtin Dep. 49:8-19, 50:8-16); GX1119 (Godfrey Dep. 52:1-53:7). It could have 25 happened more frequently if Bazaarvoice had not frustrated efforts to enable cross- 26 network syndication with PowerReviews customers in order to prevent PowerReviews 27 from offering free and open syndication of R&R to compete with Bazaarvoice. See, e.g., 28 GX711. 120 1 317. As part of its “value-based” pricing strategy, Bazaarvoice considers, among other 2 things, the volume of content and number of “nodes” a customer wishes to syndicate 3 content to when setting each customer’s individualized price. GX87 (Godfrey CID Dep. 4 102:8-103:20); cf. GX1009 (Curtin Dep. 39:19-22, 105:20-106:21). Accordingly, any 5 increase in syndication post-merger is not a cognizable, merger-specific efficiency because 6 there is insufficient evidence to show that Bazaarvoice will pass through that benefit to 7 customers instead of simply raising prices to customers by a commensurate amount. See 8 Trial Tr. 1002:4-13, 1021:20-24 (Shapiro); GX984* at 27. 9 318. Although Bazaarvoice characterizes its authentication and moderation standards and procedures as superior to those employed by PowerReviews, some former 11 United States District Court Northern District of California 10 PowerReviews customers appear to have preferred the legacy PowerReviews system. 12 GX117 (Parsons 30(b)(6) Dep. 51:15-52:18); see also, e.g., GX1000; GX999; GX998. 13 Bazaarvoice has not identified any former PowerReviews customers who complained 14 about PowerReviews’s moderation or authentication policies prior to the merger. GX117 15 (Parsons 30(b)(6) Dep. 43:19-44:9). And any PowerReviews customer that preferred the 16 Bazaarvoice moderation and authentication system had a choice and could have selected 17 Bazaarvoice as its R&R platform provider. 18 319. Prior to the merger, Bazaarvoice’s per-review moderation costs were much higher 19 than PowerReviews’. GX494 (Godfrey 30(b)(6) Dep. 42:16-45:3, 46:13-16); GX1003. As 20 a result, over time customers are likely to be charged a higher price for Bazaarvoice’s 21 moderation services, reducing any net benefit to customers. 22 320. Bazaarvoice has not shown that the merger will result in cognizable, merger- 23 specific efficiencies that will be passed through to customers and will be sufficient to 24 offset the anticompetitive effects of the transaction. Trial Tr. 923:17-24 (Shapiro). It 25 continues to innovate, as it must, but there is no evidence that the merger caused increased 26 innovation or that, absent the merger, the efficiencies and innovation claimed would not 27 have been realized. 28 121 1 DISCUSSION 2 As quoted at the outset of this Opinion, Section 7 of the Clayton Act prohibits mergers 3 whose effect “may be substantially to lessen competition, or to tend to create a monopoly.” 15 4 U.S.C. § 18. Assessing whether a merger violates Section 7 involves a burden-shifting analysis. 5 First the government must show a likelihood of anticompetitive effects. The government can do 6 this by establishing that the merger would produce a “firm controlling an undue percentage share 7 of the relevant market, and results in a significant increase in the concentration of firms in that 8 market.” Philadelphia Nat’l Bank, 374 U.S. at 363. Such a showing establishes a presumption 9 that the merger will substantially lessen competition in violation of Section 7. 10 If the government makes such a showing, the burden shifts to the defendant to rebut the United States District Court Northern District of California 11 presumption by showing why the merger is unlikely to substantially lessen competition, or by 12 discrediting the data underlying the presumption in the government’s favor. See United States v. 13 Citizens & S. Nat’l Bank, 422 U.S. 86, 120 (1975) (“the Government plainly made out a prima 14 facie case of a violation of [§] 7 under several decisions of this Court. It was thus incumbent upon 15 C&S to show that the market-share statistics gave an inaccurate account of the acquisitions’ 16 probable effects on competition”) (internal citations omitted); Philadelphia Nat’l Bank, 374 U.S. 17 at 363 (“a merger which produces a firm controlling an undue percentage share of the relevant 18 market, and results in a significant increase in the concentration of firms in that market is so 19 inherently likely to lessen competition substantially that it must be enjoined in the absence of 20 evidence clearly showing that the merger is not likely to have such anticompetitive effects”); 21 United States v. Baker Hughes Inc., 908 F.2d 981, 982-83 (D.C. Cir. 1990). 22 “If the defendant successfully rebuts the presumption of illegality, the burden of producing 23 additional evidence of anticompetitive effect shifts to the government, and merges with the 24 ultimate burden of persuasion, which remains with the government at all times.” H.J. Heinz Co., 25 246 F.3d at 715 (internal punctuation and citation omitted); Baker Hughes Inc., 908 F.2d at 983. 26 As stated at the outset of the Memorandum Opinion, the government easily established its 27 prima facie case and Bazaarvoice was unable to rebut the presumption of illegality. 28 122 1 I. THE HIGH CONCENTRATION OF THE RELEVANT MARKET IS KEY TO THIS CASE 2 3 4 5 6 7 8 9 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 In applying Section 7, courts have traditionally begun with an assessment of the relevant product and geographic markets. See, e.g., Marine Bancorporation, Inc., 418 U.S. at 618 (“Determination of the relevant product and geographic markets is “a necessary predicate” to deciding whether a merger contravenes the Clayton Act.”) (citation omitted); California v. Sutter Health Sys., 130 F. Supp. 2d 1109, 1118 (N.D. Cal. 2001) (“[t]o establish a prima facie case under Section 7 of the Clayton Act, a plaintiff must first define the relevant market”). The 2010 MERGER GUIDELINES appear to signal a turn away from the pivotal significance with which the antitrust community has traditionally imbued market definition. They state that: The Agencies’ analysis need not start with market definition. Some of the analytical tools used by the Agencies to assess competitive effects do not rely on market definition, although evaluation of competitive alternatives available to customers is always necessary at some point in the analysis. MERGER GUIDELINES § 4. The MERGER GUIDELINES nonetheless note that “Market definition can be informative regarding competitive effects.” Id. Consistent with the MERGER GUIDELINES, the government in this case asserts that “relevant markets are defined to aid the court in determining whether a defendant has the ability to exercise market power,” Govt’s Prop. Conclusion of Law ¶ 9, seemingly placing less emphasis on market definition than it was traditionally afforded. In contrast, Bazaarvoice asserts that “[t]he first element the government must establish is the relevant product market.” Bazaarvoice Prop. Conclusions of Law ¶ 22. The Court need not enter into this seemingly fertile debate over the primacy of market definition or determine whether, in this case, the government could establish anticompetitive effects without first defining a relevant market. Here the government has proposed a market consisting of R&R platforms for retailers and manufacturers. The Court agrees with that definition. Because the market for R&R platforms is highly concentrated, defining the relevant product market as R&R platforms has major consequences for this action. 28 123 1 2 A. Relevant product market The relevant product market consists of “those products to which consumers will turn, 3 given reasonable variations in price.” Lucas Auto. Eng’g, Inc. v. Bridgestone/Firestone, Inc., 275 4 F.3d 762, 767 (9th Cir. 2001). “Where an increase in the price of one product leads to an increase 5 in demand for another, both products should be included in the relevant product market.” Id. The 6 product market is “determined by the reasonable interchangeability of use or the cross-elasticity of 7 demand between the product itself and substitutes for it.” Brown Shoe, 370 U.S. at 325; Lucas 8 Auto., 275 F.3d at 767 (same); see also E. I. du Pont de Nemours & Co., 351 U.S. at 404 (product 9 market consists of products “that have reasonable interchangeability for the purposes for which 10 they are produced [based on] price, use and qualities”). United States District Court Northern District of California 11 In addition, in Brown Shoe, 370 U.S. at 325, the Supreme Court stated that “practical 12 indicia,” such as industry or public recognition of the market, the product’s peculiar characteristics 13 and uses, unique production facilities, distinct customers, distinct prices, sensitivity to price 14 changes, and specialized vendors, may aid the determination of the boundaries of the relevant 15 market. See also Olin Corp. v. F.T.C., 986 F.2d 1295, 1299 (9th Cir. 1993) (“practical indicia” 16 identified in Brown Shoe are relevant “in determining the primary market to be analyzed for 17 antitrust purposes”); Oracle Corp., 331 F. Supp. 2d at 1122 (“A relevant market may be defined 18 by reference to Brown Shoe’s ‘practical indicia.’”). 19 The evidence strongly supports the conclusion that R&R platforms are the relevant product 20 market. Bazaarvoice and PowerReviews recognized that R&R provides a unique eCommerce 21 service that cannot be replaced by other products. As Bazaarvoice CEO Collins stated during a 22 conference call with investors in November 2012, “[i]t is already well documented that authentic 23 review content is the most important content consumers rely upon to make purchase decisions, and 24 as such, we think we have a very significant and somewhat unique ability to directly influence our 25 clients’ revenue as well as aid in pricing, inventory management and product design decisions.” 26 See, e.g.,GX302 at 2. At trial, Collins testified that it will continue to be important for retailers to 27 offer R&R. Trial Tr. 752:17-20 (Collins). Similarly, PowerReviews’s CEO Comée explained that 28 R&R “was table stakes. You had to have it.” GX84 (Comée Dep. 41:5-11, 41:14-42:8). 124 1 Bazaarvoice’s recognition that PowerReviews was its primary competitor supports the 2 determination that R&R platforms are the relevant product market. Bazaarvoice’s co-founder 3 Barton wrote in an email to Facebook on June 1, 2012, that “last Thursday we announced the 4 acquisition of PowerReviews, our primary competitor.” GX507 at -412. Collins, similarly 5 testified that for “ratings and reviews, United States, PowerReviews was, I would say, our – the 6 competitor we looked at as the next best competitor.” GX81 (Collins Dep. 143:7-17). At trial, he 7 described PowerReviews as Bazaarvoice’s “primary competitor in retail.” Trial Tr. 830:19-23 8 (Collins). 9 Customers also recognized the unique value of R&R, finding R&R platforms critical in converting browsing consumers into purchasers because of the various features of the R&R 11 United States District Court Northern District of California 10 platforms that other online tools cannot provide, as discussed in Finding of Fact 92. For example, 12 Friedland testified that “[t]here is a 30 percent lift . . . if you sell product that has zero reviews 13 versus a product that has a handful of reviews. Even if they’re negative, you sell more, oddly 14 enough.” Trial Tr. 1195:7-16 (Friedland/Build.com). Other R&R customers provided similar 15 testimony. See e.g., GX76 (Waltzinger/BBBeyond Dep. 38:11-22 (Bed Bath & Beyond would be 16 at a competitive disadvantage without R&R because it is “industry standard” and “the ability for a 17 customer getting an unbiased opinion on a product relates to ultimate sales and a better experience 18 on our site”); GX193 (Walgreens Dep. 43:4-13) (“we see [R&R] as a kind of a minimum 19 requirement, to have a web site”); GX111 (Bon-Ton Dep. 44:16-45:1) (Bon-Ton had not 20 considered dropping R&R because “at this point we feel [R&R] is part of conducting our business 21 effectively”); GX141 (Golfsmith Dep. 36:12-21, 58:10-19); Trial Tr. 1218:16-22 22 (Maki/Golfsmith) (agreeing that Golfsmith needs R&R “to be a legitimate eCommerce web site”). 23 24 25 The broad array of other social commerce tools that Bazaarvoice wanted to include in the market had different purposes and were not substitutes for R&R platforms. The hypothetical monopolist test confirms that R&R platforms are the relevant product 26 market. As discussed in Findings of Fact 127-31, based on customer testimony, a SSNIP of five 27 to ten percent would not be unprofitable because few customers would abandon R&R in response 28 to such a SNIPP. See, e.g., Trial Tr. 1195:17-20 (Friedland/Build.com) (testifying that Build.com 125 1 2 would not drop R&R if its price increased by five percent). As noted in Finding of Fact 142, citing Rebel Oil, Bazaarvoice attacked Dr. Shapiro’s 3 analysis of the market for failing to consider supply substitution when defining the market. 4 Bazaarvoice overstates the holding of Rebel Oil, which teaches that supply substitution may be 5 considered in defining the market, and allocated market share, if those suppliers “can readily shift 6 their production facilities” to produce the product at issue. Rebel Oil, 51 F.3d at 1436. Rebel Oil 7 involved allegations of predatory pricing in the retail gasoline market in Las Vegas. The Ninth 8 Circuit rejected the plaintiffs’ argument that the market was limited to self-service gasoline and 9 held that the market included both self-service and full-service gasoline. The court explained that the plaintiffs’ argument “fails to account for the fact that sellers of full-serve gasoline can easily 11 United States District Court Northern District of California 10 convert their full-serve pumps, at virtually no cost, into self-serve, cash-only pumps, expanding 12 output and thus constraining any attempt by ARCO to charge supracompetitive prices for self- 13 serve gasoline.” Id. The Court concluded that “[t]he ease by which marketers can convert their 14 full-serve facilities to increase their output of self-serve gasoline requires that full-serve sales be 15 part of the relevant market. Id. 16 Rebel Oil does not stand for the proposition that it is error to define a market based only on 17 demand substitution without accounting for supply substitution. Indeed, the MERGER GUIDELINES 18 state that “[m]arket definition focuses solely on demand substitution factors.” MERGER 19 GUIDELINES § 4. Rebel Oil merely instructs that where a supplier can “easily”—i.e., “at virtually 20 no cost”—start supplying the product at issue to the relevant geographic market, that supplier 21 should be included in the market definition. Nothing in Rebel Oil states that this will necessarily 22 be the case in all mergers. In this case, based on the evidence before it, the Court concludes that 23 there are no companies likely to enter the market that warrant expanding the market definition to 24 account for supply substitution. As noted below, there are likewise no rapid entrants that should 25 be assigned market share. 26 27 28 B. Relevant geographic market The relevant geographic market is the location “where, within the area of competitive overlap, the effect of the merger on competition will be direct and immediate.” Philadelphia Nat’l 126 1 Bank, 374 U.S. at 357. The “area of effective competition in the known line of commerce must be 2 charted by careful selection of the market area in which the seller operates, and to which the 3 purchaser can practicably turn for supplies.” Id. at 359 (citation omitted). In addition, the 4 geographic market must “‘correspond to the commercial realities’ of the industry.” Brown Shoe, 5 370 U.S. at 336. Again, there was no substantial evidence to make the market broader than the United 6 7 States. Bazaarvoice and PowerReviews referred to the United States as a separate market. The 8 realities of selling and servicing the platforms, moderating content, and addressing differences in 9 SKUs explain why the market is not worldwide. 10 United States District Court Northern District of California 11 12 13 II. BAZAARVOICE CANNOT REBUT THE GOVERNMENT’S PRIMA FACIE CASE A. Bazaarvoice’s percentage share of the relevant market exceeds 50 percent and establishes a prima facie Section 7 violation In Philadelphia National Bank, the Supreme Court held that a proposed merger of 14 Philadelphia’s first and third largest banks that would result in a single bank controlling 30 percent 15 of the market share presented a threat of undue market concentration. Philadelphia Nat’l Bank, 16 374 U.S. 321 at 364 (“Without attempting to specify the smallest market share which would still 17 be considered to threaten undue concentration, we are clear that 30% presents that threat.”). 18 Notwithstanding Philadelphia National Bank, Bazaarvoice suggests that a 30 percent market share 19 does not warrant a presumption of undue concentration. In support, Bazaarvoice cites Oracle, 20 where Judge Walker concluded that “[a] presumption of anticompetitive effects from a combined 21 share of 35 percent in a differentiated products market is unwarranted.” Oracle Corp., 331 F. 22 Supp. 2d at 1123. Judge Walker reasoned that 23 24 25 In a unilateral effects case, a plaintiff is attempting to prove that the merging parties could unilaterally increase prices. Accordingly, a plaintiff must demonstrate that the merging parties would enjoy a post-merger monopoly or dominant position, at least in a “localized competition” space. 26 27 28 127 1 Id. at 1118. 17 While a 35 percent market share may not have warranted a presumption of anticompetitive 2 3 4 5 6 7 8 9 10 United States District Court Northern District of California 11 12 13 effects in that case, in this case the government established that the combined market share of Bazaarvoice and PowerReviews far exceeds 30 percent, and is in excess of 50 percent. Bazaarvoice and PowerReviews admitted as much pre-merger. Bazaarvoice’s Vice President of Strategy recognized that the R&R market was limited to Bazaarvoice and PowerReviews, stating that “My take is that there really isn’t a market for them to understand (as it relates R&R), it is us or PowerReviews and in the game of big clients: we win.” GX540 at -252. Similarly, Bazaarvoice’s founder and then-CEO, Brett Hurt, remarked that one of the benefits of acquiring PowerReviews would be “[p]otentially taking out our only competitor, . . . .” GX518 at -475. PowerReviews also recognized Bazaarvoice’s and PowerReviews’s dominance in the R&R market, repeatedly referring to the R&R market as a “duopoly” consisting or Bazaarvoice and PowerReviews. See, e.g., GX254 at 3; GX255 at -698; GX955 at -568; GX683 at -339. Dr. Shapiro analyzed the shares in two ways with data from the IR500. Dr. Shapiro 14 15 16 17 18 19 recognized that the IR500 does not provide a “perfect measure” of the R&R market, so he performed additional analyses to check the robustness of these results: he checked the results against the Fortune 500, and analyzed a random sample of actual and prospective customers identified by Dr. Shehadeh. Each analysis confirmed a combined market share in excess of 50 percent. • 20 IR500: As discussed in Findings of Fact 154 to 155, Dr. Shapiro determined that in 2012, Bazaarvoice and PowerReviews combined provided R&R to approximately 68 21 percent of the largest online retailers in North America that offered R&R, as measured 22 by the IR500. Other R&R providers in combination provided R&R to fewer than four 23 24 17 25 26 27 28 As the United States District Court for the District of Columbia noted, “[s]ome commentators have criticized this [Oracle] standard, however, because impermissible price increases can be achieved on far lower market shares than Oracle's standard evidently requires.” H & R Block, Inc., 833 F. Supp. 2d at 84 (internal punctuation omitted) (citing Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law, ¶ 914 at 84 (3d ed. 2007). The H & R Block court “decline[d] the defendants’ invitation, in reliance on Oracle, to impose a market share threshold for proving a unilateral effects claim.” H & R Block, 833 F. Supp. 2d at 85. 128 1 percent of IR500 companies that used R&R. Approximately 28 percent of IR500 2 companies that used R&R used an in-house platform. As discussed in Findings of Fact 156 to 157, Dr. Shapiro also looked to the IR500 3 4 to measure the portion of online revenues from retailers using R&R that comes from 5 retailers that use R&R provided by Bazaarvoice and PowerReviews. He concluded 6 that retailers that use R&R provided by Bazaarvoice and PowerReviews account for 56 7 percent of online revenue from retailers using R&R. Retailers using R&R provided by 8 other commercial vendors account for two percent of online revenue from retailers 9 using R&R. In-house solutions account for 42 percent of revenues from retailers using R&R, but one customer—Amazon—accounts for approximately 67 percent of that, 11 United States District Court Northern District of California 10 leading to it accounting for 28 percent of online revenue from retailers using R&R. 12 Since this second measure is based on revenue, rather than number of retailers, it is 13 useful for gauging Bazaarvoice’s and PowerReviews’s significance to brands as well as 14 retailers. The IR500 accounts for 98 percent of revenues of the thousand largest online 15 retailers as listed in the IR1000. 16 • Fortune 500: In 2012, 149 Fortune 500 companies used R&R. Bazaarvoice provided 17 R&R to 67 percent of those companies and PowerReviews provided R&R to 16 18 percent. In combination, they provided R&R to 83 percent of Fortune 500 companies 19 that used R&R. 14 percent of Fortune 500 companies that used R&R had an in-house 20 solution and three percent used another commercial supplier. See Finding of Fact 159. 21 • Dr. Shehadeh’s sample: Dr. Shapiro analyzed a random sample of 1,000 actual and 22 prospective R&R customers that Dr. Shehadeh selected from a pool of 29,494. The 23 1,000 accounts were not limited to brands or retailers or big or small entities. After 24 limiting the data set to domestic retailers and brands that have a R&R platform, Dr. 25 Shapiro found that Bazaarvoice had a market share of 38 percent and PowerReviews 26 had a market share of 21 percent. See Finding of Fact 160. Bazaarvoice and 27 PowerReviews accordingly provided R&R to 59 percentof the companies in Dr. 28 Shehadeh’s random sample that used R&R. 40 percent of those companies used an in129 house solution and two percent used another commercial provider. 1 The Court recognizes that the above measures do not perfectly capture the combined 2 3 entity’s share of the R&R market. Nonetheless, each of the measures reveals the same 4 same basic market structure: that Bazaarvoice and PowerReviews are the two dominant providers 5 of R&R and they have a combined market share in excess of 50 percent. Bazaarvoice criticizes the market share proposed by the government because it does not 7 assign market share to rapid entrants and expanders. Bazaarvoice argues that Amazon, Google, 8 Oracle, and other companies should be assigned market shares as rapid entrants. Bazaarvoice is 9 correct that rapid entrants may properly be assigned market shares. As the MERGER GUIDELINES 10 state, “[f]irms that would rapidly and easily enter the market in response to a SSNIP are market 11 United States District Court Northern District of California 6 participants and may be assigned market shares.” MERGER GUIDELINES § 9. However, as stated in 12 Findings of Fact 222 to 239, the Court finds that in this case there are no firms that qualify as rapid 13 entrants that should be assigned market share. Dr. Shapiro employed the Herfindahl–Hirschmann Index (“HHI”), a measure of market 14 15 concentration often employed by Courts in Section 7 cases and endorsed by the Merger 16 Guidelines. 18 See, e.g., H.J. Heinz Co., 246 F.3d at 716 (“Market concentration, or the lack 17 thereof, is often measured by the Herfindahl–Hirschmann Index”); Merger Guidelines § 5.3. 18 Under the Merger Guidelines, a market with an HHI above 2500 is considered a “highly 19 concentrated market.” The Merger Guidelines further state that 20 Mergers resulting in highly concentrated markets that involve an increase in the HHI of between 100 points and 200 points potentially raise significant competitive concerns and often warrant scrutiny. Mergers resulting in highly concentrated markets that involve an increase in the HHI of more than 200 points will be presumed to be likely to enhance market power. The presumption may be rebutted by persuasive evidence showing that the merger is unlikely to 21 22 23 24 25 26 27 28 18 The MERGER GUIDELINES are not binding on the courts, but they “are often used as persuasive authority when deciding if a particular acquisition violates anti-trust laws.” Chicago Bridge, 534 F.3d at 432; see also United States v. Kinder, 64 F.3d 757, 771 & n.22 (2d Cir. 1995) (Leval, J., dissenting) (“Although it is widely acknowledged that the Merger Guidelines do not bind the judiciary in determining whether to sanction a corporate merger or acquisition for anticompetitive effect, courts commonly cite them as a benchmark of legality.”) (listing cases). 130 1 2 3 4 5 6 7 8 9 10 United States District Court Northern District of California 11 12 13 14 15 16 17 18 19 20 21 22 23 24 enhance market power. Based on customer revenues, Dr. Shapiro calculated the pre-merger HHI at 2674 and postmerger at 3915. Dr. Shapiro also calculated the HHI based on customer count, which revealed a pre-merger HHI of 2365 and a post-merger HHI of 4590. By all measurements presented, whether based on market share or market concentration, the government easily made a prima facie showing of a Section 7 violation. B. Bazaarvoice failed to rebut the government’s prima facie case “The more compelling the prima facie case, the more evidence the defendant must present to rebut it successfully.” Baker Hughes Inc., 908 F.2d at 991. A variety of factors, none of which is dispositive, can rebut a presumptive showing of a Section 7 violation. The factors include low barriers to entry, actual entry, weakness of the data underlying the prima facie case, industry structure, cross-elasticities of demand and supply, product differentiation, and efficiency. See, e.g., United States v. Syufy Enterprises, 903 F.2d 659, 664 (9th Cir. 1990) (low barriers and actual meaningful entry rebutted prima facie case); Baker Hughes Inc., 908 F.2d at 985-86 (low barriers to entry, combined with “flawed underpinnings” of the government’s prima facie case and the sophistication of the relevant market’s consumers, rebutted government’s prima facie case). Firms are more likely to enter a market where there are no significant barriers to entry. H & R Block, Inc., 833 F. Supp. 2d at 73 (“Determining whether there is ease of entry hinges upon an analysis of barriers to new firms entering the market or existing firms expanding into new regions of the market.”) (citation omitted). The presence or absence of significant barriers to entry are therefore frequently “crucial considerations in a rebuttal analysis.” Baker Hughes Inc., 908 F.2d at 987; see also Syufy, 903 F.2d at 664 (“If there are no significant barriers to entry, however, eliminating competitors will not enable the survivors to reap a monopoly profit; any attempt to raise prices above the competitive level will lure into the market new competitors able and willing 25 26 27 28 131 1 to offer their commercial goods or personal services for less.”). 19 i. 2 Prospective entry Bazaarvoice argued at trial that there are no barriers to entry preventing expansion by 3 4 existing firms or entry by new firms to replace the competition previously provided by 5 PowerReviews. In contrast, the government argued that significant barriers to entry, including 6 network effects from syndication, high switching costs, reputation, and technical capabilities 7 prevent firms outside the market from meaningfully entering the market and inhibit existing 8 suppliers’ ability to expand to compete with Bazaarvoice. The prospect that other firms will enter the market can stimulate competition and rebut a 9 prima facie violation of Section 7 only “if such entry will deter or counteract any competitive 11 United States District Court Northern District of California 10 effects of concern so the merger will not substantially harm customers.” MERGER GUIDELINES § 9; 12 see also Rebel Oil, 51 F. 3d at 1440 (same). Entry must be “timely, likely, and sufficient in its 13 magnitude, character, and scope to deter or counteract the competitive effects of concern.” 14 MERGER GUIDELINES § 9; see also Chicago Bridge, 534 F.3d at 429-30 (“courts have generally 15 concluded that for entry to constrain supracompetitive prices, the entry has to be of a ‘sufficient 16 scale’ adequate to constrain prices and break entry barriers”). 17 Since the government has established its prima facie case, Bazaarvoice has the burden of 18 showing that low barriers to entry rebut the likelihood that the merger will have anticompetitive 19 effects. See, e.g., H & R Block, Inc., 833 F. Supp. 2d at 73. Here, Bazaarvoice identified Google 20 and Facebook, among others, as potential entrants into the R&R market. As discussed in detail in 21 22 23 24 25 26 27 28 19 Bazaarvoice asserts that the “consensus view” is that two years is the proper “time frame for entry and competitive effects analysis.” Bazaarvoice Post-Trial Brief at 18. Dr. Shapiro and Dr. Shehadeh likewise stated that the time frame is around two years. Tr. 2035:20-22 (Shehadeh) (“So I would think that if we don't see effects in the next two years, then we don't need to worry about effects farther out in the future.”); Tr. 1023:25-1025:2 (Shapiro) (“So the time frame for entry, I'm thinking about a year or two, and these entry barriers are just not easy to surmount.”). The MERGER GUIDELINES state that entry is timely if it is “rapid enough to make unprofitable overall the actions causing those effects and thus leading to entry, even though those actions would be profitable until entry takes effect.” MERGER GUIDELINES § 9.1. The Court agrees that two years is an appropriate time-frame in this case. Entry within two years is likely to undo the anticompetitive effects created by the merger such that the merger would be unprofitable, whereas entry beyond two years is not. 132 1 Findings of Fact 239-264, the Court finds that syndication, switching costs, intellectual 2 property/know how, and reputation are formidable barriers to new firms entering the market for 3 R&R platforms and to existing R&R providers expanding their operations to replace the 4 competition previously provided by PowerReviews. 5 In addition, Bazaarvoice gave no reason why those firms were likely to enter the market. 6 There was no evidence that any company had made even preliminary analyses of the viability of 7 joining the market. Indeed, companies like Google and Facebook are seen as complementary to 8 the R&R market at present, and there are no indications that this will change in the foreseeable 9 future. It may be that if Bazaarvoice eventually moves past the R&R market to big data and advertising, it will be in direct competition with Google and Amazon. But that is not the situation 11 United States District Court Northern District of California 10 it faces with the R&R market today. Moreover, the mere fact that Bazaarvoice, Amazon, Google 12 and others all operate in the broader eCommerce space, does not mean that they are all 13 competitors for antitrust purposes and should be placed in the same product market. See, e.g., 14 F.T.C. v. Staples, Inc., 970 F. Supp. 1066, 1075 (D.D.C. 1997) (“the mere fact that a firm may be 15 termed a competitor in the overall marketplace does not necessarily require that it be included in 16 the relevant product market for antitrust purposes”). 17 The marketplace may be filled with many strong and able companies in adjacent spaces. 18 But that does not mean that entry barriers become irrelevant or are somehow more easily 19 overcome. To conclude otherwise would give eCommerce companies carte blanche to violate the 20 antitrust laws with impunity with the excuse that Google, Amazon, Facebook, or any other 21 successful technology company stands ready to restore competition to any highly concentrated 22 market. As a matter of economic reality, companies do not simply enter any market they can— 23 they will only do so if it is within their strategy to do so and they have the requisite ability to do 24 so. The fact that social commerce and eCommerce tastes and products are developing and 25 constantly changing does not diminish the applicability of the antitrust laws—they apply in full 26 force in any market. There is no antitrust exemption that allows the market-leading company in a 27 highly concentrated market to buy its closest competitor, even within the evolving social 28 commerce space, when the effect is likely to be anticompetitive. 133 Bazaarvoice cites Baker Hughes, 908 F.2d at 988-89 for the proposition that it need not 1 2 “present actual examples of firms that are ‘poised for future expansion’” because “a firm that 3 never enters a given market can nevertheless exert competitive pressure on that market.” 4 Bazaarvoice Prop. Conclusions of Law ¶ 67. It is not clear to the Court that this assertion requires 5 an analysis distinct from the analysis regarding prospective entry above. In any event, the 6 principle in Baker Hughes was limited to situations where “barriers to entry are insignificant.” 20 7 Baker Hughes, 908 F.2d at 988. As noted in Findings of Fact 240-265, barriers to entry are not 8 insignificant in this case. To the contrary, the barriers to entry are formidable. Bazaarvoice has 9 been in the market for eight years, yet there is almost no evidence of significant entry of a commercial competitor other than PowerReviews. Given the significant to barriers to entry, the 11 United States District Court Northern District of California 10 Court finds that there is no outside firm that exerts sufficient competitive pressure to ameliorate 12 the showing of likely anticompetitive effect of the merger. ii. 13 Actual entry In post-merger cases, evidence of actual entry of new firms which renders the merged 14 15 entity unable to maintain its market share can rebut a prima facie showing. See, e.g., Syufy, 903 16 F.2d at 665. As with prospective entry, the actual entry must be of a scale sufficient to counteract 17 any anticompetitive effect that results from the challenged merger. The Ninth Circuit explained 18 that “[t]he fact that entry has occurred does not necessarily preclude the existence of ‘significant’ 19 entry barriers. If the output or capacity of the new entrant is insufficient to take significant 20 business away from the predator, they are unlikely to represent a challenge to the predator’s 21 market power.” Rebel Oil, 51 F.3d at 1440; cf Philadelphia Nat’l Bank, 374 U.S. 321, 367 n.43 22 (“The test of a competitive market is not only whether small competitors flourish but also whether 23 consumers are well served.”). For example, in Rebel Oil, the Ninth Circuit found that the entry of “two small rivals” did 24 25 26 27 28 20 The complete quotation from Baker Hughes reads: “Furthermore, the supposed ‘quick and effective’ entry requirement overlooks the point that a firm that never enters a given market can nevertheless exert competitive pressure on that market. If barriers to entry are insignificant, the threat of entry can stimulate competition in a concentrated market, regardless of whether entry ever occurs.” Baker Hughes, 908 F.2d at 988 (emphasis added). 134 1 not contradict or render unreasonable the plaintiffs’ argument that there were significant entry 2 barriers. Rebel Oil, 51 F.3d at 1441. In contrast, in Syufy, the actual entry of new firms rebutted 3 the presumption of a Section 7 violation because the entry was of a scale that prevented the 4 defendant from maintaining its market share. Syufy, 903 F.2d at 665 (“Las Vegas’ first-run film 5 market was more competitive when this case came to trial than before Syufy bought out Mann, 6 Plitt and Cragin.”). 21 While a few companies have entered the market recently, their entry is of such a minimal 7 scale that it is not close today, and is unlikely to be close in the next two years, to replacing 9 PowerReviews. Bazaarvoice identified third parties, including Reevoo, Pluck, Gigya, Lithium, 10 and Rating-Systems, as competitors of Bazaarvoice that constrain Bazaarvoice’s anticompetitive 11 United States District Court Northern District of California 8 power in the R&R market. However, as discussed in detail in Findings of Fact 185 to 208, none 12 of those firms provides significant competition to Bazaarvoice and they are “unlikely to represent 13 a challenge to the [Bazaarvoice’s] market power.” Rebel Oil, 51 F.3d at 1440. iii. 14 Merger specific efficiencies will not overcome anticompetitive harm 15 As noted in Findings of Fact 310 to 320, the Court is not persuaded that Bazaarvoice’s 16 acquisition of PowerReviews resulted or will result in efficiencies that overcome the merger’s 17 anticompetitive harms. Bazaarvoice expects only limited cost savings from the acquisition. The 18 merger has provided Bazaarvoice new opportunities for content syndication by increasing the 19 amount of data that Bazaarvoice controls and the number of customers that Bazaarvoice 20 syndicates to and from. But the ability to syndicate R&R content between PowerReviews and 21 Bazaarvoice customers—“cross network syndication”—is not a merger specific efficiency. Such 22 syndication was feasible and was occurring for some customers prior to the merger. In addition, there is no evidence that Bazaarvoice’s proffered “architectural improvements” 23 24 could not have been completed without the acquisition of PowerReviews. To the contrary, 25 Bazaarvoice had begun developing the improvements at issue prior to its acquisition of 26 27 28 21 As stated in more detail below, post-merger evidence of actual entry is less susceptible to manipulation by the parties than other types of post-merger evidence, such as lack of price increases or continued innovation, and is therefore more probative. 135 1 2 PowerReviews and expected to implement them successfully irrespective of the acquisition. While Bazaarvoice has increased its R&D spending since the acquisition, this appears 3 driven by Bazaarvoice’s need to continue to innovate in the dynamic eCommerce market, rather 4 than by its acquisition of PowerReviews. 5 iv. 6 The probative value of post-merger evidence, or lack thereof Because this case involves a consummated merger, the Court must determine how much weight to give post-merger evidence. A court may consider post-merger evidence “diminish[ing] 8 the probability or impact of anticompetitive effects,” but the probative value of such evidence is 9 “extremely limited.” United States v. General Dynamics Corp., 415 U.S. 486, 504 (1974) (citing 10 E. I. duPont de Nemours & Co., 353 U.S. at 597; United States v. Continental Can Co., 378 U.S. 11 United States District Court Northern District of California 7 at 463)). The Supreme Court has explained that “[t]he need for such a limitation is obvious. If a 12 demonstration that no anticompetitive effects had occurred at the time of trial or of judgment 13 constituted a permissible defense to a [Section] 7 divestiture suit, violators could stave off such 14 actions merely by refraining from aggressive or anticompetitive behavior when such a suit was 15 threatened or pending.” General Dynamics Corp., 415 U.S. at 504-05. Moreover, “the essential 16 question remains whether the probability of such future impact exists at the time of trial.” Id. at 17 505 (emphasis added). The MERGER GUIDELINES likewise state that “a consummated merger may 18 be anticompetitive even if such effects have not yet been observed, perhaps because the merged 19 firm may be aware of the possibility of post-merger antitrust review and moderating its conduct.” 20 MERGER GUIDELINES § 2.1.1. 21 Consistent with General Dynamics Corp., the probative value of post-acquisition evidence 22 is particularly limited “whenever such evidence could arguably be subject to manipulation.” 23 Chicago Bridge, 534 F.3d at 435 (emphasis in original). This is especially true when the parties 24 are aware of the government’s scrutiny and the potential for a court challenge. If a court 25 incorrectly relies on post-merger testimony that a merged entity has not raised prices and the court 26 blesses the transaction, there is little to prevent the merged entity from creating anticompetitive 27 effects at a later time. 28 Conversely, post-merger evidence that is not subject to manipulation is of greater probative 136 1 value, and may be dispositive. See, e.g., United States v. Int’l Harvester Co., 564 F.2d 769, 779- 2 80 (7th Cir. 1977) (“the value of post-acquisition evidence is greater where the evidence is beyond 3 the power of the parties to manipulate) (internal quotation omitted). For example, post-merger 4 evidence of significant changes in the relevant market can rebut a government’s prima facie case 5 of a Section 7 violation. In General Dynamics Corp., the Supreme Court approved of the district 6 court’s reliance on “evidence relating to changes in the patterns and structure of the coal industry 7 and in United Electric’s coal reserve situation after the time of acquisition in 1959.” 415 U.S. 486, 8 501. The Court explained that “[s]uch evidence could not reflect a positive decision on the part of 9 the merged companies to deliberately but temporarily refrain from anticompetitive actions, nor could it reasonably be thought to reflect less active competition than that which might have 11 United States District Court Northern District of California 10 occurred had there not been an acquisition in 1959.” Id. The Court concluded in light of these 12 “these fundamental changes in the structure of the market for coal, the District Court was justified 13 in viewing the statistics relied on by the Government as insufficient to sustain its case.” Id. 14 Post-merger evidence of actual entry that has prevented the merged entity from 15 maintaining its market share can also rebut a prima facie case. For example, in Syufy the Ninth 16 Circuit affirmed the district court’s finding that post-merger evidence showed that actual entry 17 prevented the merged entity from achieving anticompetitive market power. In the present case, 18 however, there is no evidence of changes in the R&R industry that call into question the 19 government’s prima facie case. There is some evidence that providers of eCommerce platforms 20 are moving towards providing “bundles” or “suites” of eCommerce products, including R&R. But 21 there is no evidence that this trend has fundamentally altered the R&R market in a way that 22 undermines the government’s prima facie case. Nor is there evidence of actual meaningful entry 23 that has prevented Bazaarvoice from maintaining its post-acquisition market share. 24 The complexity of the economic and legal issues raised in antitrust actions warrants 25 affording limited value to lay testimony regarding the effects of a merger. In Philadelphia 26 National Bank, the Supreme Court discounted competitors’ pre-merger testimony that competition 27 would continue to be vigorous after the proposed merger. The Court explained that this “lay 28 evidence on so complex an economic-legal problem as the substantiality of the effect of this 137 1 merger upon competition was entitled to little weight, in view of the witnesses’ failure to give 2 concrete reasons for their conclusions.” Philadelphia National Bank, 374 U.S. at 367. The Court 3 finds that, in this case, post-merger customer testimony regarding the effect of the merger upon 4 competition is also entitled to limited weight given the customers’ narrow perspective on the 5 economic and legal questions before the Court and their lack of visibility into Bazaarvoice’s 6 practices. In addition, as noted above, the customers were not privy to most of the evidence 7 presented to the Court, including that of the economic experts; many of the customers had paid 8 little or no attention to the merger; and each had an idiosyncratic understanding of R&R based on 9 the priorities of their company and their different levels of knowledge, sophistication, and 10 United States District Court Northern District of California 11 experience. v. 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Syufy does not compel an “alternative methodology” for assessing Section 7 cases involving consummated mergers or otherwise warrant judgment for Bazaarvoice Bazaarvoice argues that in Syufy, 903 F.2d at 665-69, the Ninth Circuit articulated an “alternative methodology” for assessing Section 7 cases involving consummated mergers “that can obviate the need to analyze all the many factors in the traditional burden-shifting approach” used in pre-merger cases. Bazaarvoice argues that under this methodology, “if evidence post-merger shows (1) that price competition and innovation in the affected product market continue at the same level, and/or (2) that affected customers have testified the merger is not harmful, and/or (3) that entry is easy, a violation of Section 7 of the Clayton Act has not been established and the traditional approach can be bypassed.” Bazaarvoice concludes that “the government has failed to satisfy the requirements of Syufy, thereby warranting judgment in favor of the defendant.” The Court does not agree that Syufy signals an “alternative methodology” for analyzing post-merger Section 7 cases or that the traditional burden-shifting approach can be bypassed if the post-merger evidence identified by Bazaarvoice is present. With particular relevance to this case, the Court does not agree that the government cannot carry its burden if post-merger evidence shows continued price competition and innovation or if “affected customers have testified the merger is not harmful.” In Syufy, the Ninth Circuit looked at the post-merger landscape and determined that 138 1 Syufy’s acquisitions did not injure competition. But the dispositive factor in Syufy was the lack of 2 meaningful barriers to entry, which enabled the actual entry of new competitors. As the Ninth 3 Circuit stated 4 The government concedes that there are no structural barriers to entry into the market: Syufy does not operate a bank or similar enterprise where entry is limited by government regulation or licensing requirements. Nor is this the type of industry, like heavy manufacturing or mining, which requires onerous front-end investments that might deter competition from all but the hardiest and most financially secure investors. Nor do we have here a business dependent on a scarce commodity, control over which might give the incumbent a substantial structural advantage. Nor is there a network of exclusive contracts or distribution arrangements designed to lock out potential competitors. To the contrary, the record discloses a rough-and-tumble industry, marked by easy market access, fluid relationships with distributors, an ample and continuous supply of product, and a healthy and growing demand. It would be difficult to design a market less susceptible to monopolization. 5 6 7 8 9 10 United States District Court Northern District of California 11 12 Syufy, 903 F.2d at 666-67 (9th Cir. 1990) (internal citation omitted). By the time of trial—indeed, 13 shortly after the acquisitions at issue—new competitors had already entered the market, preventing 14 Syufy from maintaining the market share briefly obtained through its acquisitions. Cf. id. at 66515 66 (“In evaluating monopoly power, it is not market share that counts, but the ability to maintain 16 market share.”) (emphasis in original). Reviewing the evidence of post-acquisition competition, 17 the Ninth Circuit concluded that “[i]t all boils down to this: Syufy’s acquisitions did not short 18 circuit the operation of the natural market forces; Las Vegas’ first-run film market was more 19 competitive when this case came to trial than before Syufy bought out Mann, Plitt and Cragin.” 20 Id. at 665. 21 Syufy did not overturn the long-established methodology for reviewing Section 7 claims, 22 which applies equally to pre- and post-acquisition cases. To the extent that Syufy found post23 acquisition evidence of insignificant barriers to entry and actual entry of new competitors 24 dispositive, that is consistent with the traditional burden-shifting methodology for analyzing 25 section 7 cases, employed since before the enactment of the Hart-Scott-Rodino Act of 1976 (the 26 “HSR Act”) limited the number of Section 7 cases involving consummated, as opposed to 27 proposed, mergers. 28 139 The HSR Act requires entities larger than specified thresholds to file various notifications 1 2 with the government before merging. 22 15 U.S.C. § 18a. The HSR Act thus enables the Antitrust 3 Division of the United States Department of Justice and the Federal Trade Commission to review 4 proposed mergers before they are consummated and to seek to enjoin the mergers under Section 7 5 of the Clayton Act. As a consequence, most Section 7 actions over the last four decades, 6 including many frequently-cited decisions, involve proposed mergers, not consummated mergers. 7 See, e.g., H & R Block, Inc., 833 F. Supp. 2d at 51-52; Oracle Corp., 331 F. Supp. 2d at 1112; 8 Baker Hughes Inc., 908 F.2d at 987. But the relative dearth of post-merger Section 7 cases over the last 40 years does not mean 9 that the Court is not bound by long-standing and controlling authority regarding the proper 11 United States District Court Northern District of California 10 framework for analyzing Section 7 cases or the probative value of post-merger evidence. Rather, 12 Supreme Court authority predating the enactment of the HSR Act establishes and affirms the 13 burden-shifting framework for analyzing Section 7 cases and applies equally to pre- and post- 14 merger cases. See, e.g., General Dynamics Corp., 415 U.S. 486 (post-merger); Philadelphia Nat’l 15 Bank, 422 U.S. 86 (pre-merger). Supreme Court authority predating the HSR Act likewise 16 establishes the framework for evaluating post-merger evidence. See, e.g., General Dynamics 17 Corp., 415 U.S. 486. Neither the enactment of the HSR Act nor the passage of time has overruled 18 or superseded this authority. 19 III. APPLICATION OF ANTITRUST LAW TO DYNAMIC MARKETS This case inescapably adds fuel to the debate over the proper role of antitrust law in rapidly 20 21 changing high-tech markets. Some maintain that antitrust law is ill-suited to these dynamic 22 markets, arguing, for example, that it undermines high-tech innovation (see, e.g., Robert J. Barro, 23 Why the Antitrust Cops Should Lay Off High-Tech, BUSINESS WEEK, Aug. 17, 1998, at 20 (“the 24 best policy for the government in the computer industry is to stay out of it”)) or that market power 25 is transitory in high-tech industries where competitive ideas can overcome entrenchment. Others 26 argue that antitrust law is well-suited to address the challenges presented by the hi-tech economy 27 22 28 Merging entities smaller than the reporting thresholds are not required to provide pre-merger notifications. That was the case here. 140 1 (see, e.g., Richard Posner, “Antitrust in the New Economy,” 68 ANTITRUST L.J. 925 (2001) 2 (“antitrust doctrine is supple enough, and its commitment to economic rationality strong enough, 3 to take in stride the competitive issues presented by the new economy”)) or, as the Ninth Circuit 4 noted 35 years ago, that “rapid technological progress may provide a climate favorable to 5 increased concentration of market power rather than the opposite.” Greyhound Computer Corp., 6 Inc. v. Int’l Bus. Machines Corp., 559 F.2d 488, 497 (9th Cir. 1977). 7 It is not the Court’s role to weigh in on this debate. The Court’s mission is to assess the 8 alleged antitrust violations presented, irrespective of the dynamism of the market at issue. See, 9 e.g., United States v. Microsoft Corp., 253 F.3d 34, 49-50 (D.C. Cir. 2001) (noting that whether or not Microsoft was correct that it operated in a market where “entrenchment may be temporary, 11 United States District Court Northern District of California 10 because innovation may alter the field altogether . . . . does not appreciably alter our mission in 12 assessing the alleged antitrust violations in the present case”). As the Court has set forth in detail, 13 while Bazaarvoice indisputably operates in a dynamic and evolving field, it did not present 14 evidence that the evolving nature of the market itself precludes the merger’s likely anticompetitive 15 effects. 16 CONCLUSION 17 Bazaarvoice violated Section 7 of the Clayton Act by purchasing its closest and only 18 serious competitor, creating the likelihood of an anticompetitive effect in the R&R market in the 19 United States. The parties shall file a Joint Statement Regarding Remedy Phase on January 17, 20 2014, and attend a Case Management Conference on January 22, 2014, at 9 a.m. in Courtroom 2, 21 450 Golden Gate Avenue, San Francisco, California, to discuss the remedy phase of this case. 22 23 24 25 26 IT IS SO ORDERED. Dated: January 8, 2014 ______________________________________ WILLIAM H. ORRICK United States District Judge 27 28 141

Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.

Why Is My Information Online?