CNG Financial Corporation v. Google Inc

Filing 74

RESPONSE in Opposition re 69 MOTION to Exclude Testimony of Plaintiff's Expert Michael Mazis filed by Plaintiff CNG Financial Corporation, Counter Defendant CNG Financial Corporation. (Attachments: # 1 Affidavit /Declaration of Hunter in support of CNG Opposition to Motion# 2 Exhibit Exhibit A to Attachment# 3 Exhibit Exhibit B to Attachment to Opposition# 4 Exhibit Exhibit C to Attachment to Opposition# 5 Exhibit Exhibit D to Attachment to Opposition# 6 Exhibit Exhibit D - Part B to Attachment to Opposition# 7 Exhibit Exhibit E to Attachment to Opposition to Motion to Exclude) (Hunter, Barry)

CNG Financial Corporation v. Google Inc

Doc. 74

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 1 of 11

UNITED STATES DISTRICT COURT SOUTHERN DISTRICT OF OHIO WESTERN DIVISION
CNG FINANCIAL CORPORATION
Plaintiff/CounterclaimDefendant,
vs.
Case No. 1 :06-cv-040

Chief Judge Sandra S. Beckwith Magistrate Timothy S. Black

GOOGLE INC.
Defendant/Counterclaim- :

CNG FINANCIAL CORPORATION'S OPPOSITION TO GOOGLE'S MOTION TO EXCLUDE TESTIMONY OF PLAINTIFF'S

EXPERT MICHAEL MAZIS

Plaintiff.
For its Opposition to the Defendant Google, Inc. ("Google")'s Motion to Exclude

Testimony of Plaintiffs Expert Michael Mazis, CNG Financial Corporation ("CNG") states as
follows:

INTRODUCTION
The common theme-the only theme, really-running through Google's motion to
exclude the testimony of CNG's expert is that Dr. Mazis "tested the wrong issue," and that his

testimony is therefore irrelevant. Google's Motion to Exclude Testimony of Plaintiffs Expert
Michael Mazis ("Google's Motion to Exclude Mazis") at 7. Google's self-proclaimed

characterization of the correct issue is "whether Google's use of CNG's mark as a keyword

trgger causes consumer confusion," and the sole means of testing this issue, according to

Google, is to utilize a generic search term as a control. However, contrary to what Google

asserts in its Motion, CNG's complaint in this case is not solely about the use of its trademark as

an AdW ords keyword; it is about the confusion being engendered by the appearance of
competitor's sponsored links when a consumer searches the Check N Go trademark on Google's

Dockets.Justia.com

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 2 of 11

search engine.l Google's own surey expert has acknowledged this point. Hunter Decl., ii 2,

Exhibit A (Simonson depo., pp. 129:17-129:22). In other words, this Court should not view
consumer confusion in a vacuum. Part of the issue is, and always has been, that the sponsored

links triggered by the use of Check N Go as a keyword are coined by CNG's questionable
competitors, and are by design vague, imprecise, and thus confusing. Thus, and as explained

more fully below, Google's assertion that the surey control must isolate the confusion
engendered by the use of the keyword from the confusion engendered by the appearance of the
sponsored links misapprehends CNG's claim in this case.

Moreover, Google's characterization of Dr. Mazis' survey as simply testing whether
payday loan sponsored links generate more confusion than insurance or penny stock sponsored
links is nothing more than rhetorical flourish.2 The survey was designed to see if the appearance

of competitors' sponsored links in response to the use of Check N Go as a keyword generate a

material amount of confusion. The control was designed to deduct from these results the
potential effect of "noise" and other causes for confusion-for example, the ability to show any

sponsored links on the generated Check N Go results page-about which CNG does not
complain.

3 Dr. Mazis wanted to test the confusion generated by Google's practice of selling

CNG's trademark to CNG's competitors, and his surey was designed to and did accomplish that

objective. His testimony will therefore assist this court in its understanding of the issue of
i See CNG's Complaint, ir 29 ("By the operation and design of its AdWords Program, Google causes consumers

who specifically intend and desire to find a Check 'N Go website to be directed, instead, to a search result page Check 'N Go Online and the Check 'N Go Companies."). competitors of containg links to the websites of
2 Essentially any surey with a control can be simlarly trvialized as "merely showing" that one tye of

practice (the

test group) causes more confusion than another tye of practice (the control group). For example, if 75% of
respondents given a new aspirin product reported gettng signficantly better relief from headache symptoms than they do from aspirn while 10% ofrespondents given a placebo reported the same thig, you could trvialize the test as "merely showing" that the new product was better than sugar water at relieving headache symptoms. In fact, what ths study really showed was that a net 65% of the respondents reported that the new pil was very effective at
curing headaches.

3 Google does not dispute that this is the appropriate purose of a surey control.

2

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 3 of 11

consumer confusion generated by Google's sale of CNG's trademark to CNG's competitors, and
should not be excluded.

Finally, it is notable that Google's latest Motion simply repeats the same challenge to Dr.

Mazis' survey that it asserted at sumary judgment-without ever addressing or purporting to

refute the detailed point-by-point response to that challenge set forth in CNG's prior filed
summar judgment opposition papers. CNG suggests that the failure by Google's able counsel
to squarely address these points, which are largely repeated in this opposition, serves to
underscore their validity.

ARGUMENT
A. That Google does not agree with CNG's premise on what this case is about does not

make it irrelevant under the federal rules or Daubert-Dr. Mazis' opinion wil assist this Court in understanding the evidence and in determining a fact in issue.
As an initial matter, it is important to remember that this case wil be tried by this Court,

not by a jury. Sixth Circuit cours, as well as those in other jursdictions, have emphasized that

the so-called "gatekeeper" fuction of Fed. R. Evid. 702 is less significant in the context of a
bench tral. See, e.g., Deal v. Hamilton Co. Bd. of Educ., 392 F.3d 840, 852 (6th Cir. 2004)

("The 'gatekeeper' doctrne was designed to protect juries and is largely irrelevant in the context

of a bench triaL"); Whirlpool Props., Inc. v. LG Elecs. Us.A., Inc., No. 1:03-cv-4l4, 2006 WL
62846, at *1 n.l (W.D. Mich. Jan. 10,2006) ("The Sixth Circuit recognzes that the gatekeeping
function is less critical in the context of a bench triaL"); see also Us. v. Brown, 415 F.3d 1257,
1268 (11 th Cir. 2005) ("(The traditional barrers to opinion testimony) are even more relaxed in a

bench trial situation, where the judge is serving as factfinder... There is less need for the
gatekeeper to keep the gate when the gatekeeper is keeping the gate only for himself."); Gibbs v.

Gibbs, 210 F.3d 491, 500 (5th Cir. 2000) ("most of the safeguards provided for in Daubert are

3

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 4 of 11

not as essential in a case such as this where a district judge sits as the trier of fact in place of a

jur") .
Nevertheless, even if the gatekeeper doctrine applies here in full force, CNG respectfully
submits that the testimony of Dr. Mazis will do much more, and certainly no less, than assist this

Court in its understanding of the issue of consumer confusion. In its efforts to exclude the
testimony of Dr. Mazis, Google's only arguent is that Dr. Mazis' opinion is not relevant.4
Examining relevance in the context of the admission of expert testimony includes analyzing
whether the testimony "fits" the facts of the case and whether the testimony addresses a subject

matter on which the factfinder can be assisted by an expert. Daubert v. Merrell Dow Pharms.,
Inc., 509 U.S. 579, 591 (1993); see also Ullman v. Auto-Owners Mut. Ins. Co., No. 2:05-cv-

1000, 2007 WL 1858146, at *3 (S.D. Ohio June 26, 2007) ("The second prong of the gatekeeping role requires an analysis of whether the expert's reasoning or methodology can be

properly applied to the facts at issue, that is, whether the opinion is relevant to the facts at

issue."). In Daubert, the Supreme Cour notes that relevant evidence is defined elsewhere in the
Federal Rules of Evidence as "that which has any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would

be without the evidence." 509 U.S. at 587 (quoting Fed. R. Evid. 402). The Court goes on to
provide that "(t)he Rules' basic standard of

relevance thus is a liberal one." Id.

Google asserts that Dr. Mazis' "surey is not relevant to this litigation because...it does

not consider whether Google's use of CNG's mark as a keyword trigger causes consumer
confusion." Google's Motion to Exclude Mazis at 6. As discussed above, Google's narrow

definition of what is and is not relevant confusion fails to encompass that confusion created by
4 Google does not address, and thus apparently does not object to, the qualifications of Dr. Mazis, the suffciency of
his facts or data, or the reliability of

his methods, the other factors contemplated by Fed. R. Evid. 702.

4

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 5 of 11

the use of CNG's mark as a keyword trggering competitive and confusingly similar sponsored
links.5 Consideration of the confusion created in the actual context of the sponsored links

generated when inputting the Check N Go keyword will certainly assist this Cour in determining

a factual dispute in this case-the likelihood of consumer confusion when competitive sponsored
links are trggered by the input of Check N Go as a search term.

Google's dispute with Dr. Mazis' factual bases for his opinion, if to be credited by this
cour at all, goes only to the evidentiary weight afforded his opinion, not its admissibility. See,
e.g., Daubert, 509 U.S. at 595-96 (gatekeeper analysis should focus on "principles and

methodology, not on the conclusions they generate"); AHP Subsidiary Holding Co. v. Stuart
Hale Co., 1 F.3d 611, 618 (7th Cir. 1993) ("(w)hile there wil be occasions when the proffered
survey is so flawed as to be completely unhelpful to the trer of fact and therefore inadmissible,
such situations wil be rare,,).6 A cour's gatekeeper role under the federal rules is not meant to

5 As "support" for its mischaracterization of Dr. Mazis' surey and the relevance of

what it tested, Google attempts

to distort Dr. Mazis' deposition testimony, quoting him entirely out of context as stating that he didn't test the confsion engendered by the use of Check N Go as an AdWords keyword. However, what Dr. Mazis was explainng was that his test was for confsion generated not just by the entry of Check N Go as a keyword, but also
by competitive sponsored lins appearig in response to the entry of the Check N Go keyword, as ilustrated in this exchange during Dr. Mazis' deposition:

Q. The purose of the study was to tr to measure the effect of "Check 'n
Go" as a search term correct? As opposed to in the text of ads? Isn't that what you said earlier?

A. Yes, in part. Right.
Q. Well, then, what's the other part?
A. Well, the other part has to do with-It's when you-when "Check 'n

Go" is placed in the search window, when the sponsored lin-when the results page come up and there's sponsored lins on there, the question is to what extent
are people confsed as to the source of those sponsored lins. So, I mean,

there's two elements. One is entering "Check 'n Go" in the search window or
search box, and the second is the sponsored lins, the competitive sponsored

lins.
Hunter Decl., ir 3, Exhbit B (Mazis depo., p. 17:6-17:22).

6 See also Prudential Ins. Co. v. Gibraltar Fin. Corp., 694 F.2d 1150, 1156 (9th Cir. 1982) ("Techncal uneliability
goes to the weight accorded a surey, not its admissibility."); Leelanau Wine Cellars, Ltd. v. Black & Red, Inc., 452

F. Supp. 2d 772, 778 (W.D. Mich. 2006) ("...flaws in surey methodology go to the evidentiary weight of the surey rather than its admssibility"); Whirlpool Props., 2006 WL 62846, at *8 (same); McCarty § 32:170
(improper surey may be excludable "in extreme case.. .But the majority rule is that while technical deficiencies can reduce a surey's weight, they will not prevent the surey from being admitted into evidence."). Google also cites

5

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 6 of 11

replace the adversary system. Indeed, given the liberal thrst of the federal rules, especially in
the context of a bench trial, Google's objections to Dr. Mazis' testimony are more suitable for

demonstration durng its cross-examination of Dr. Mazis and through presentation of its own
contrary evi ence.

'd 7

Simply because Google disagrees with CNG's determination of the facts

pertinent to the determination of consumer confusion in the context of this case does not render
the opinion ofCNG's expert irrelevant or unhelpful to this Court in its role as factfinder.

B. Because CNG's complaint in this case is not solely about the use of its trademark as

an AdWords keyword-but also about the appearance of the sponsored links
triggered by a Google search of that trademark-Dr. Mazis did not need to select a
survey control which would isolate the effects of the one from the other.

As explained above, Dr. Mazis tested, on a control group, whether the respondents
believed that sponsored links advertising the services of insurance companies and stock brokers

which appeared in response to their Google search of Check N Go were sponsored by Check N

Go. Dr. Mazis then netted out, from the confusion results reported from his test group (the
legally relevant confusion), the confusion results reported from his control group (the legally
irrelevant confusion). The difference yielded the net confusion rate.

Google argues that a more proper control would have been to test for the amount of
confusion experienced by respondents searching for generic keywords such as "payday loans."

Of course, Google can cite to no case in which a survey is excluded simply because a more

to AutoZone, Inc. v. Tandy Corp., 373 F.3d 786 (6th Cir. 2004), for the proposition that Sixth Circuit cours wil exclude a surey that provides "no evidence of actual confsion." In AutoZone, however, the surey at issue was conducted to consider the strength of the mark, not confsion. The cour, considering the survey in the context of its examiation of the eight factors relevant for likelihood of confsion, found that the surey did not provide evidence
of actual confsion, not a surrising rinding considerig the surey was not designed to so provide.

7 See, e.g., Daubert, 509 U.S. at 596 ("Vigorous cross-examiation, presentation of contrary evidence, and careful

instrction on the burden of proof are the traditional and appropriate means of attackig shaky but admssible
evidence.").

6

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 7 of 11

effective control might have been selected.8 Even sureys with no controls, like the surey that

Google unsuccessfully challenged in the American Blind case for having failed to isolate the

confusion caused by the use of the keyword from the confusion caused by the appearance of
sponsored links (precisely the same wrongheaded challenge it levels here) have routinely been
admitted.9 Moreover, in any event, Google's suggested control makes no sense in this case.
The basis for Google's arguent here is a sentence in Judge Brinkema's opinion in

Government Employees Insurance Co. v. Google, Inc., 2005 WL 1903128, at *5 (E.D. Va.)

("GEICO "), discussing what Judge Brinkema concluded, after admitting the survey at trial,

were methodological flaws in GEICO's surey in that case. Id. ("As a threshold matter, the
control retained the use of "GEICO" as a keyword, which itself was alleged to be a source of

confusion"). However, critical distinctions between the facts before Judge Brinkema and the
facts of

this case make her observations about the control in that case wholly inapplicable here.

In the GEICO survey, unlike in Dr. Mazis's survey, respondents were shown search
pages with the term GEICO appearing in the title ofthe sponsored links. Thus, it was difficult to

determine whether the reported confusion resulted from the use of GEICO as a keyword or,
8 Google cites no case, and CNG has not uncovered any, where a surey was wholly excluded for utilizing a flawed
control. To the contrary, cours have often admtted sureys despite having an arguably flawed control, see

Merisant Co. v. McNeil Nutritionals, LLC, 2007 U.S. Dist. LEXIS 27681, at *25-*32 (E.D. Pa. April 12, 2007)

(par's alleged "improper" choice of control "go to the weight to be accorded to the surey rather than to its
admssibility"). The two cases Google cites as examples of cours that excluded a surey on the basis of inadequate
surey controls-Cumberland Packing Corp. v. Monsanto Co., 32 F. Supp. 2d 561, 574 (E.D.N.Y. 1999) and Hil's

Pet Nutrition, Inc. v. Nutro Prods., Inc., 258 F. Supp. 2d 1197, 1210 (D. Kan. 2003)-did not actually do so. The cours in both were ruling on plaintiffs' motion for a prelimary injunction, and considered the sureys in analyzing evidence of actual confsion. In Cumberland Packing, the cour found the surey "unpersuasive" based on its improper unverse, umealistic market conditions, and problems with the control. In Hil's Pet Nutrition, considering
a surey offered to show implied falsity for a false advertsing claim, the cour discounted the survey on reliability

grounds, including issues with the control, biased questions, and insignificant findings. Neither cour, however, was considering a motion to wholly exclude the testiony of the expert at triaL.
9 Hunter Decl. ir 4, Exhbit C, p. 15; see also Ironclad, L.P. v. Poly-America, Inc., 2000 U.S. Dist. LEXIS I0728, at

*24 (N.D. Tex. July 28, 2000) ("However, the Cour need not exclude the surey due to the lack of control, as

generally, techncal deficiencies go to the weight rather than admissibility. In light of the deficiency, the Cour
discounts the weight of

the survey, yet still gives Mr. Stone's results some consideration.").

7

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 8 of 11

instead, from the use of GEICO in the title of the sponsored links. Because, by the time of tral,

Google had disabled all of the sponsored links containing the term GEICO in their text, while
continuing to allow advertisers to use GEICO as a keyword trigger, the confusion caused by the

appearance of GEICO in the text of the sponsored links shown to the surey respondents was
legally irrelevant to GEICO's injunction claim and needed to be isolated from the surey results.
The entirety of

the surey criticisms in the GEICO opinion related to the inability to separate out

the legally relevant confusion engendered by the use of GEICO as a keyword from the legally
irrelevant confusion engendered by the appearance of GEICO in the text of the sponsored links.
And the only way to do that, ifGEICO was not removed from the text of

the sponsored link, was
10

to remove GEICO from the keyword search and then to test the remaining confusion.

Here, unlike in GEICO, Dr. Mazis's surey respondents were not shown sponsored links

with Check N Go in the title; to the contrary, they were only shown links containing terms that
are permissible under Google's current policies and that continue to appear on the search pages
trggered by a search of the keyword Check N GO.11 Thus, unlike in the GEICO survey, the

confusion engendered by the appearance of the sponsored links shown to Dr. Mazis's surey
respondents was legally relevant (since those links were representative of the links that are stil
being triggered by the use of Check N Go as a keyword), and Dr. Mazis did not need to isolate

the confusion engendered by the use of Check N Go as a keyword from the confusion
10 Moreover, simply because the GEICO cour endorsed a partcular surey design does not signfy that this design

is the only one that wil produce relevant results, especially considering the easily distiguished facts. Cf Whirlpool Props., 2006 WL 62846, at *36-*37 (assertng that prior case cited by plaintiff "did not establish a surey template that must be followed in every case," and that defendant's surey design provided "suffciently reliable and relevant informtion to pass the gatekeeping theshold").

11 In fact, by avoiding the use of many of the more deceptively titled lins, that are permtted under Google's
policies and that continue to appear today, e.g., "Check in Go" and "CheckiGo.com," Dr. Mazis's surey tended to underreport the likely measure oflegally relevant confsion. And because the sponsored lins shown to Dr. Mazis's respondents contained no text suggesting any connection with Check N Go, there is no reason to believe that any respondents would have believed that those listings were affiliated with Check N Go had they appreciable amount of been shown those lin in response to their search of a generic keyword such as "payday loans."

8

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 9 of 11

engendered by the appearance of the sponsored links on the Check N Go search page. Thus,
absent the need to separate out these two types of confusion, which was necessitated under the
unique facts in GEICO, there was no need for Dr. Mazis to control for the confusion engendered
12
by a search of a generic keyword.

Finally, for all its grping about CNG's failure to use a generic search term, such as

"payday loans," as a control and for all of its claims that the need for such control was so
obvious, Google's own experts in GEICO, both in the initial survey and expert report and in the

rebuttal expert report, did not use and seemingly never even considered the use of a generic
search term as a control.13 Further, as emphasized in CNG's summary judgment opposition

papers, Google did not bother to conduct its own surey utilizing its preferred control to refute
Dr. Mazis' findings, a detail courts in the Sixth Circuit have noted "significantly undermines the
effcacy" of any argument to exclude another pary's expert.14

12 Google distorts Dr. Mazis' testimony on ths point as well. Rather than testifying that he ignored the GEICO
cour's criticism of the GEICO plaintiffs surey for not using a different keyword as a control, what Dr. Mazis

actually said was that that criticism did not apply to this case. Hunter Decl, ir 3, Exhbit B (Mazis depo. p. 58:358: 14).

13 In its initial surey report in GElCO, Google's first surey expert, Dr. Jacoby, utilized several controls; however,

none was a generic search term related to GEICO's insurance field. Hunter Decl, ir 5, Exhbit D. In its rebuttal
expert report, Google's second surey expert, Dr. Hamar Simonson, criticized the control used by GEICO's expert,

but he did not discuss the apparently now-obvious vires of using a generic search term related to insurance as a control. /d., ir 6, Exhbit E.
14 On this point, Google's statement about the significance CNG assigns to Google not conducting its own surey is

utterly disingenuous. According to Google, CNG argued that "because Google's expert criticized CNG's surey but did not conduct one himself, CNG's survey must be relevant and inadmissible." (CNG assumes this last part is a
tyographical error, as CNG certainy did not argue its surey must be inadmssible.) Google then proceeds to take

its self-embellished version of CNG's argument to "its logical extreme"-essentially comparing Dr. Mazis' detailed surey of consumer confusion to a study proving "clowns are fuy." Google's Motion to Exclude Mazis at 7 n.2. What CNG actually stated in its opposition to Google's motion for summry judgment is that Google's failure to present its own surey evidence challenging Dr. Mazis' findings is telling and "compels against excluding Dr. Mazis' study." See CNG's Memorandum in Opposition to Google's Motion for Summary Judgment at 21. CNG did not then and does not now represent that the absence of competing survey evidence, on its own, warrants the its surey. However, Google's flippant analogies and disregard for what CNG and cours have actually admssion of asserted cannot explain away Sixth Circuit law, also cited by CNG in its opposition papers, supporting that the absence of competig surey evidence "signifcantly undermines the effcacy of plaintiffs' arguments that defendants' evidence is so uneliable that it must be excluded." Whirlpool, 2006 U.S. Dist. LEXIS 1378, at *l 1*12.

9

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 10 of 11

CONCLUSION

For the reasons stated above, Google's Motion to Exclude Testimony of Plaintiffs
Expert Dr. Michael Mazis should be denied.

Respectfully submitted,

Isl Bary D. Hunter An Gallagher Schoen Frost Brown Todd, LLC 2200 PNC Center 201 East Fifth Street Cincinnati, OH 45202-4182
TeL.: [phone]
Fax: [phone]

Bary D. Hunter

Medrth Lee Norman Frost Brown Todd LLC 250 West Main Street, Suite 2700 Lexington, Kentucky 40507
Tel. [phone]
Fax: [phone]

Attorneys for CNG FINANCIAL CORPORATION

10

.'

Case 1:06-cv-00040-SSB-TSB

Document 74

Filed 07/20/2007

Page 11 of 11

CERTIFICATE OF SERVICE
I certify that a copy of the foregoing was electronically fied with the Clerk of the Cours
using the CMlECF, system which wil send notification of such fiing to the following, this 20th
day of July, 2007:

Kenneth F. Seibel, Esq. Jacobs, Kleinman, Seibel & McNally
2300 Kroger Building

1014 Vine Street Cincinnati, OH 45202

Michael H. Page, Esq. Klaus H. Ham, Esq. Keker & VanNest LLP 710 Sansome Street San Francisco, CA 94107

Isl Barv D. Hunter Attorney for Plaintiff/Counter Defendant

11
LEXLibrary 0102393.0533475 339659vl