Shenwick v. Twitter, Inc. et al
Filing
113
ORDER GRANTING IN PART AND DENYING IN PART DEFENDANTS' MOTION TO DISMISS by Judge Jon S. Tigar granting in part and denying in part 91 Motion to Dismiss. (wsn, COURT STAFF) (Filed on 10/16/2017)
1
2
3
4
UNITED STATES DISTRICT COURT
5
NORTHERN DISTRICT OF CALIFORNIA
6
7
DORIS SHENWICK, et al.,
Plaintiffs,
8
v.
9
10
TWITTER, INC., et al.,
Defendants.
11
United States District Court
Northern District of California
Case No.16-cv-05314-JST
ORDER GRANTING IN PART AND
DENYING IN PART DEFENDANTS’
MOTION TO DISMISS
Re: ECF No. 91
12
Twitter is a social media company that, like other social media companies, depends on
13
14
advertising revenue. Accordingly, how many users Twitter has, and whether those users are
15
engaged with Twitter‘s content, are deeply important to its success as a company. In their
16
complaint, Plaintiffs allege that Twitter executives knowingly made inaccurate public statements
17
regarding these metrics, and failed to disclose internal information about them, resulting in an
18
inflated share price that fell when the truth about user engagement became known.
Defendants now move to dismiss Plaintiffs‘ complaint. ECF No. 91. The Court concludes
19
20
that Plaintiffs‘ claims are, for the most part, adequately pleaded. The Court will grant the motion
21
in part and deny it in part.
22
I.
REQUESTS FOR JUDICIAL NOTICE
23
The Court first addresses Defendants‘ requests for judicial notice. ECF Nos. 92, 106. ―As
24
a general rule, [the Court] may not consider any material beyond the pleadings in ruling on a Rule
25
12(b)(6) motion.‖ United States v. Corinthian Colleges, 655 F. 3d 984, 998–99 (9th Cir. 2011)
26
(internal quotation marks and citations omitted). However, ―[t]he [C]ourt may judicially notice a
27
fact that is not subject to reasonable dispute because it: (1) is generally known within the trial
28
court‘s territorial jurisdiction; or (2) can be accurately and readily determined from sources whose
1
accuracy cannot reasonably be questioned.‖ Fed. R. Evid. 201(b). The Court ―must take judicial
2
notice if a party requests it and the court is supplied with the necessary information.‖ Fed. R.
3
Evid. 201(c).
Alternatively, under the incorporation by reference doctrine, a court may ―take into
4
account documents whose contents are alleged in a complaint and whose authenticity no party
6
questions, but which are not physically attached [to] the [plaintiff‘s] pleading.‖ Knievel v. ESPN,
7
393 F. 3d 1068, 1076 (9th Cir. 2005) (quoting In re Silicon Graphics Inc. Sec. Litig., 183 F. 3d
8
970, 986 (9th Cir. 1999)). The Court may consider the entire document, even if only portions
9
were quoted or referenced in the Complaint. See Bell Atl. Corp. v. Twombly, 550 U.S. 544, 569,
10
n. 13 (2007) (citing Fed. R. Evid. 201) (holding that courts are ―entitled to take notice of the full
11
United States District Court
Northern District of California
5
contents of the published articles referenced in the complaint, from which the truncated quotations
12
were drawn.‖).
First Request for Judicial Notice – ECF No. 92
13
A.
14
First, Defendants request that the Court take judicial notice of the following documents,
15
which Plaintiff references in her Complaint: (1) Twitter‘s Registration Statement, ECF No. 92-1;
16
(2) Twitter‘s Form 10-Q for Q4 2014, ECF No. 92-4; (3) Twitter‘s Form 10-Q for Q1 2015, ECF
17
No. 92-5; (4) a transcript of Twitter‘s October 27, 2014 Earnings Call for Q3 2014,1 ECF No. 92-
18
6; (5) a transcript of Twitter‘s February 5, 2015 Earnings Call for Q4 2014, ECF No. 92-7; (6) a
19
transcript of Twitter‘s April 28, 2015 Earnings Call for Q1 2015, ECF No. 92-8; (7) a transcript of
20
Twitter‘s July 28, 2015 Earnings Call for Q2 2015, ECF No. 92-9; (8) PowerPoint slides presented
21
by Twitter at its ―Analyst Day‖ investor conference on November 12, 2014, ECF No. 92-10; (9)
22
Nick Bilton, Twitter is Betting Everything on Jack Dorsey, VANITY FAIR (June 1, 2016), ECF
23
No. 92-13; (10) a letter from Twitter to the SEC‘s division of Corporation Finance, ECF No. 92-
24
15; (11) Zachary Edward, How a Small Change by Apple Cost Twitter Millions of Users,
25
QUARTZ (February 5, 2015), ECF No. 92-16; (12) Jim Edwards, Twitter‘s user growth problem
26
27
28
1
This stands for the third quarter of 2014. The Court uses this type of abbreviation throughout the
Order.
2
1
may be worse than you think, BUSINESS INSIDER (February 6, 2015), ECF No. 92-17; and (13)
2
Lara O‘Reilly, Twitter admits 8.5% of its users are bots, MARKETING WEEK (August 12,
3
2014), ECF No. 92-18. ECF No. 92 at 2-5.
The set of documents listed above are referenced in the Complaint and Plaintiff does not
4
5
question their authenticity. ECF No. 96 at 2, 5. Thus, they are properly subject to judicial notice.
6
Id. at 5. However, because the facts contained in the documents are subject to reasonable dispute,
7
―the Court takes judicial notice only of the statements contained therein, but not for the purpose of
8
determining the truth of those statements.‖ Shaev v. Baker, No. 16-CV-05541-JST, 2017 WL
9
1735573, at *7 (N.D. Cal. May 4, 2017) (quoting In re LDK Solar Sec. Litig., 584 F. Supp. 2d
1230, 1254 (N.D. Cal. 2008).2 Accordingly, the Court takes judicial notice of ECF Nos. 92-1, 92-
11
United States District Court
Northern District of California
10
4, 92-5, 92-6, 92-7, 92-8, 92-9, 92-10, 92-13, 92-15, 92-16, 92-17, and 92-18.
12
Second, Defendants request judicial notice of the following SEC filings and media reports
13
that were not referenced in the Complaint: (1) Twitter‘s Form 10-Q for Q2 2014, ECF No. 92-2;
14
(2) Exhibit C, Twitter‘s Form 10-Q for Q3 2014, ECF No. 92-3 (3) Jillian D‘Onfro, Twitters
15
Admits 5% Of Its ‗Users‘ Are Fake, BUSINESS INSIDER (October 3, 2013), ECF No. 92-11; and
16
(4) Todd Spangler, Twitter‘s Spam Headache: More than 10 Mil Accounts Might Be Bogus,
17
VARIETY (October 4, 2013), ECF No. 92-12. ECF No. 92 at 2-4.
Both the SEC filings and the news articles are judicially noticeable to show what was
18
19
available to the market. See Metzler Inv. GMBH v. Corinthian Colleges, Inc., 540 F. 3d 1049,
20
1064 n. 7 (9th Cir. 2008) (holding that the district court properly took judicial notice of publicly
21
available financial documents and SEC filings); Von Saher v. Norton Simon Museum of Art at
22
Pasadena, 592 F. 3d 954, 960 (9th Cir. 2009) (internal quotation marks omitted) (holding that
23
courts ―may take judicial notice of publications introduced to indicate what was in the public
24
realm at the time, not whether the contents of those articles were in fact true‖). The Court
25
therefore takes judicial notice of ECF Nos. 92-2, 92-3, 92-11, and 92-12, without determining the
26
2
27
28
It is not disputed that Defendants actually made the statements contained in these documents, but
that does not mean the Court may assume that the contents of these documents ―are true for
purposes of a motion to dismiss under Rule 12(b)(6).‖ Compare United States v. Ritchie, 342
F.3d 903, 908 (9th Cir. 2003).
3
1
truth of the statements therein.
Third, Defendants request that the Court take judicial notice of Defendant Noto‘s Form 4.
2
3
ECF No. 92-14. Plaintiff opposes this request. ECF No. 92 at 8. The Court acknowledges that it
4
has taken judicial notice of SEC Forms 4 on more than one occasion. Curry v. Yelp Inc., No. 14-
5
CV-03547-JST, 2015 WL 1849037, at *4 (N.D. Cal. Apr. 21, 2015); Sec. & Exch. Comm'n v.
6
Bardman, 216 F. Supp. 3d 1041, 1058 (N.D. Cal. 2016). In both cases, however, the request for
7
judicial notice was unopposed. Moreover, in most cases where courts have taken judicial notice of
8
Forms 4, the plaintiff had alleged insider trading by the defendant, or claimed that financial motive
9
established scienter. E.g., City of Royal Oak Ret. Sys. v. Juniper Networks, Inc., 880 F. Supp. 2d
1045, 1059 (N.D. Cal. 2012) (taking judicial notice of Forms 4 because ―plaintiffs‘ scienter and
11
United States District Court
Northern District of California
10
insider trading allegations do rely expressly on [defendants‘] stock sales‖); Morgan v. AXT, Inc.,
12
No. C 04-4362 MJJ, 2005 WL 2347125, at *12 (N.D. Cal. Sept. 23, 2005) (taking judicial notice
13
of the defendant‘s Forms 4 after the plaintiff referenced his stock sales in the complaint);
14
Wietschner v. Monterey Pasta Co., 294 F. Supp. 2d 1102, 1109 (N.D. Cal. 2003) (taking judicial
15
notice of Forms 4 because they were ―integral to the stock sale allegations made in the
16
Complaint‖); but see Westley v. Oclaro, Inc., 897 F. Supp. 2d 902, 928 (N.D. Cal. 2012) (taking
17
judicial notice of Forms 4 even through ―the SAC does not rely on insider trading allegations to
18
demonstrate scienter‖). Where plaintiffs have not raised the issue of stock sales in the complaint,
19
some courts have declined to take judicial notice of Forms 4. E.g., In re Adaptive Broadband Sec.
20
Litig., No. C 01-1092 SC, 2002 WL 989478, at *20 (N.D. Cal. Apr. 2, 2002); Maiman v. Talbott,
21
No. SACV090012AGANX, 2010 WL 11421950, at *7 (C.D. Cal. Aug. 9, 2010) (joining courts
22
that decline to take judicial notice of Forms 4). Therefore, it is noteworthy that here, Plaintiff did
23
not discuss stock sales in the Complaint, and she did not argue that Defendants‘ financial motives
24
establish scienter. Moreover, the fact of the stock purchases is only relevant insofar as the Court
25
draws the inference that the purchases negate scienter. Although the Court may ―compare
26
competing3 explanations and inferences,‖ those inferences still must ―arise from the facts alleged
27
3
28
As Plaintiff explains, there are other potential explanations for Defendant Noto's stock purchases.
ECF No. 94 at 38-39.
4
1
in the complaint.‖ In re MRV Commc‘ns, Inc. Derivative Litig., No. CV 08-03800 GAF RCX,
2
2010 WL 5313442, at *4 (C.D. Cal. Dec. 27, 2010) (citing Tellabs, Inc. v. Makor Issues & Rights,
3
Ltd., 551 U.S. 308 (2007). The Complaint mentions no stock sales of any kind. For all these
4
reasons, the Court will not take judicial notice of Defendant Noto‘s Forms 4.
B.
5
Second Request for Judicial Notice – ECF No. 106
Defendants filed a second request for judicial notice in connection with their reply in
6
support of the motion to dismiss. ECF No. 106. The documents identified in the request are: (1)
8
Twitter‘s Form 10-Q for Q1 2014, ECF No. 106-1; (2) Twitter‘s Form 10-Q for Q2 2015, ECF
9
No. 106-2; (3) Twitter‘s Form 10-Q for Q3 2015, ECF No. 106-3; (4) Twitter‘s Form 10-Q for Q1
10
2016, ECF No. 106-4; (5), Twitter‘s Form 10-Q for Q2 2016, ECF No. 106-5; (6) Twitter‘s Form
11
United States District Court
Northern District of California
7
10-Q for Q3 2016, ECF No. 106-6; (7) a transcript of Twitter‘s July 29, 2014 Earnings Call for Q2
12
2014, ECF No. 106-7; (8) a transcript of Twitter‘s October 27, 2015 Earnings Call for Q3 2015,
13
ECF No. 106-8; (9) a transcript of Twitter‘s statements at a March 3, 2015 investor conference
14
hosted by Morgan Stanley, ECF No. 106-9; (10) a transcript of a November 6, 2013 speech given
15
by then SEC Chairperson May Jo White, ECF No. 106-10; (11) Facebook Inc.‘s 2014 Form 10-K,
16
ECF No. 106-11; (12) Facebook Inc.‘s Form 10-Q for Q1 2015, ECF No. 106-11; and (13) Snap
17
Inc.‘s Form 10-Q for Q1 2017, ECF No. 106-13. Three of these documents, ECF Nos. 106-7,
18
106-9, 106-10, are referenced in the Complaint. ECF No. 106 at 3-4.
Like the SEC filings that were the subject of Twitter‘s first request for judicial notice, the
19
20
documents listed above are noticeable either under the incorporation by reference doctrine, or
21
because SEC filings are public documents. Metzler Inv., 540 F. 3d at 1064 n. 7 (holding that the
22
district court properly took judicial notice of publicly available financial documents and SEC
23
filings). The Court grants the request.
24
II.
INTRODUCTION
25
This is a securities class action on behalf of all persons who purchased or otherwise
26
acquired Twitter common stock between February 6, 2015 and July 28, 2015, inclusive (the ―Class
27
Period‖), against Twitter and certain of its officers and/or directors for violations of §§ 10 and
28
20(a) of the Securities Exchange Act of 1934 (the ―Act‖). ECF No. 1 (―Compl.‖) ¶ 2. Generally,
5
1
Plaintiff Doris Shenwick (―Plaintiff‖) alleges that ―Twitter and certain of its officers and/or
2
directors [] made materially false and misleading statements during the Class Period in press
3
releases and filings with the SEC and in oral statements to the media, securities analysts and
4
investors.‖ Id.4
Defendant Twitter (―Twitter,‖ or ―the Company‖) is a ―global social media platform.‖ Id.
5
6
¶ 13. Defendant Costolo served as Chief Executive Officer (―CEO‖) of Twitter during the Class
7
Period until his resignation on July 1, 2015. Id. ¶ 14. Defendant Noto served as Chief Financial
8
Officer (―CFO‖) of Twitter during the class period. Id. ¶ 15. Both Costolo and Noto were
9
―control persons‖ within the meaning of the Act.
Twitter’s Metrics
A.
11
United States District Court
Northern District of California
10
Twitter uses three principal metrics to measure ―its financial health and growth prospects‖:
12
(1) ―monthly active users or ―MAU‖ (the number of users on the platform in a given month),‖
13
(2) ―those users‘ daily activity (user engagement),‖ and (3) ―advertising engagements (the ability
14
of the Company to turn user activity into advertising revenue).‖ Id. ¶ 20. These three metrics are
15
highly ―interrelated.‖ Id. ¶ 21. First, ―[i]nformation about Twitter‘s user engagement is essential
16
to understanding its MAU growth prospects.‖ Id. ¶ 22. ―As new users become harder to find and
17
MAU growth slows, user engagement helps increase user retention and reduce [user attrition]
18
which has a direct impact on MAU growth.‖ Id. Put another way, ―[s]tagnant DAU5 growth will
19
eventually cause MAU growth to stall.‖ Id. Second, user engagement drives advertising
20
engagements: ―[t]he more often users are on the platform to view advertisements (the more
21
22
23
24
25
26
27
28
4
At the motion to dismiss phase, the Court accepts as true all material facts alleged in the
complaint, together with reasonable inferences to be drawn from those facts. Navarro v. Block,
250 F.3d 729, 732 (9th Cir. 2001).
5
DAU stands for ―Daily Active Users,‖ and includes any user who accesses the Twitter platform
on a given day. ―‗DAU‘ and ‗DAU metric‘ . . . refer generally to DAU as a measure of user
engagement. This includes the various formats in which DAU was expressed internally and
externally by Twitter. For example, at Analyst Day, the DAU metric was expressed as a ratio of
the number of daily active users compared to the number of monthly active users in Twitter‘s top
twenty markets (referred to as the DAU/MAU ratio).‖ Id. at 12 n.10. Other social media
companies also disclose DAU. Id. ¶ 124.
6
1
engaged users are), the higher Twitter‘s advertising revenues.‖ E.g., id. ¶ 23. Twitter
2
acknowledged this connection in its SEC filings. Id. ¶ 85(g) (―User growth trends reflected in the
3
number of MAUs [and] user engagement trends . . . are key factors that affect our revenue.‖)6
This interplay means that ―neither MAU nor user engagement provides a complete picture
4
of the business‖ without the other. Id. ¶ 24. For example, Twitter ―could have millions of users
6
sign up (i.e., high MAU) but if those users logged in to the platform only once a month (i.e., low
7
engagement), the business would suffer because fewer ads can be sold.‖ Id. Twitter confirmed
8
this connection in its 2013 Registration Statement, explaining that ―[t]o the extent our user growth
9
rate slows, our success will become increasingly dependent on our ability to increase levels of user
10
engagement.‖ Id. Former Twitter employees serving as confidential witnesses (―CWs‖)7 likewise
11
United States District Court
Northern District of California
5
described the low utility of MAU, particularly in predicting advertising revenue, absent
12
accompanying data on user engagement. Id. ¶ 75 (CW-98) (―MAU says nothing about frequency‖
13
of use and therefore does not provide ―sufficient detail to measure an advertising opportunity.‖);
14
Id. ¶ 76 (CW-10) (―‗MAU was a terrible metric‘ and there is no way you could ‗judge the health
15
of the company using MAU alone.‘‖). Tech analysts also expressed reservations about relying on
16
MAU as a meaningful predictor without additional information about user engagement. For
17
example, Jan Dawson, chief analyst and founder of the tech research firm Jackdaw explained
18
MAU‘s drawbacks:
19
Monthly usage metrics [MAUs] indicate very little about true engagement on a
platform, because using an app every 30 days isn‘t that much different from never
using it at all. For social and communication apps, the key is daily usage because it
tells you how people are really engaging.
20
21
22
6
23
24
At the hearing on this motion, Defendants‘ counsel acknowledged that user growth and user
engagement were two of the most important indicators of the financial health of a social media
company. ECF No. 112 at 8-9.
7
25
26
27
28
These CWs are referred to in this Order according to their numbering in the Complaint: CW-1,
CW-2, etc. The Court does not address Defendants‘ supplemental ―appendix‖ cataloguing their
objections to the CW testimony, which should have been included in the body of Defendants‘
motion.
8
Twitter objects that CW-9 (and several others) was not employed during the Class Period. ECF
No. 91-1 at 13. But the timeframe of CW-9‘s employment has little bearing on whether he can
provide relevant testimony about the interaction between DAU and MAU as a general matter.
7
1
Id. ¶ 121 (internal alterations omitted). Finally, ―Twitter‘s direct peers consistently reported both
2
total users and user engagement.‖ Id. ¶ 121 (listing Facebook, LinkedIn, Snapchat, Instagram, and
3
others).
4
B.
Pre-Class Period Activity
5
Until the beginning of the Class Period, Twitter measured user engagement using ―timeline
6
views,‖ which were a ―proxy for the amount of content [Twitter‘s] users consume.‖ Id. ¶¶ 20, 36.
7
In mid-2014, Twitter reported a ―trend of declining timeline views,‖ which it claimed was driven
8
by ―changes‖ to Twitter‘s platform that ―allow[ed] users to more efficiently access [Twitter‘s]
9
content.‖ Id. ¶ 26 (2Q 2014 Earnings Call, 3Q 2014 Earnings Call). Twitter told investors that
―the same product improvements that had resulted in fewer timeline views ultimately would lead
11
United States District Court
Northern District of California
10
to more satisfied users, which in turn would lead to higher user engagement and MAU growth.‖
12
Id. In other words, Twitter ―reassured investors‖ that there was nothing to worry about in terms of
13
user growth and engagement. Id. New users, Noto claimed in 2014, were ―just as engaged‖ as
14
existing users. Id. ¶ 27 (Q1 2014 Earnings Call).
15
In November 2014, ―Defendants held an all-day pep rally with analysts (‗Analyst Day‘) to
16
promote their vision to reinvigorate Twitter‘s user growth.‖ Id. ¶ 3. At Analyst Day, Noto and
17
Costolo announced that Twitter would double its users (from 284 million MAU to 550 million
18
MAU) in the intermediate term and reach over a billion users in the longer term. Id. ¶ 29. To
19
achieve this MAU goal, Twitter planned to ――build an engaging experience . . . to have those users
20
be engaged [and] stay engaged.‖ Id. ¶ 30. Specifically, Twitter committed to the ―operational
21
goal of building the world‘s largest daily audience‖ by ―increasing engagement and improving
22
retention.‖ Id. ¶ 31. The Company‘s DAU to MAU ratio was described as one of its ―major
23
growth drivers,‖ and Twitter tied a projected $500 million in revenue growth to a 3% increase in
24
the ratio in its top 20 markets – from 48% to 51%. Id. ¶¶ 39, 85(j). Twitter‘s stock price rose 6%
25
in response to these statements at Analyst Day. Id. ¶ 32.
26
27
28
After Analyst Day, Costolo repeated the importance of user engagement to investors and
analysts, even more so than MAU:
[Q:] Should people still focus on monthly active users? Is that what you look at on
8
1
2
the inside, is that the number-one metric you want to drive growth on?
[A:] Our number-one priority is to strengthen the core and make Twitter an
increasingly daily use case for the people who come to Twitter . . . .
3
Id. ¶ 33 (February 5, 2015 Business Insider Interview with Costolo). The content of analysts‘
4
reporting on Twitter‘s Analyst Day demonstrated that they got the message. One summarized as
5
follows: ―TWTR set some ambitious goals as it aims to have the largest DAU in [the] world with
6
over $14 billion of annual revenue.‖ Id. ¶ 39.
Within the company, the way Twitter measured user engagement was DAU. For example,
8
―CW-1 reported that DAU was the primary engagement metric that Twitter tracked internally after
9
Timeline Views were no longer being reported.‖ Id. ¶ 67. CW-2 claimed that DAU ―was talked
10
about on a daily basis by Twitter employees.‖ Id. ¶ 68; see also id. ¶ 71 (―[T]he most important
11
United States District Court
Northern District of California
7
metrics to Twitter were MAU and DAU, which were monitored closely by Twitter‘s
12
management.‖). In sum, as Defendants later acknowledged, ―[t]he best way to quantify the impact
13
of engagement . . . was . . . DAU to MAU.‖ Id. ¶ 85(b).
14
C.
Class Period Activity
15
Despite a heavy internal focus on DAU and the Company‘s publicly stated goal of
16
increasing daily use, Twitter did not ―tell investors that DAU was its primary user engagement
17
metric‖ or ―provide meaningful updates on user engagement trends, in particular DAU.‖ Id. ¶ 40.
18
Instead, throughout the Class Period, Twitter kept its user engagement statements positive but
19
vague, and highlighted MAU growth.
20
On February 5, 2015, Twitter filed a Form 8-K with the SEC containing a press release
21
announcing the Company‘s fourth quarter and fiscal year 2014 financial results, and held an
22
Earnings Call to explain the quarterly results. Id. ¶ 79. Costolo acknowledged lower than
23
expected MAU growth at the end of 2014, but described signs of a rebound in early 2015:
24
We ended the quarter with 288 million monthly active users. We added 4 million
users this quarter and 47 million across 2014. There are quarter-specific factors
that impacted our net adds in Q4, which includes seasonality and a couple issues
related to the launch of iOS 8. We'll discuss that in more detail later in this call.
Importantly, I want to highlight that the user numbers we saw in January of this
year indicate that our MAU trend has already turned around and our Q1 trend is
likely to be back in the range of absolute net adds that we saw during the first three
quarters of 2014.
25
26
27
28
9
1
Id. ¶ 88; ECF No. 92-7 at 3.9 When asked about the claims of ―acceleration‖ in MAU growth,
2
Costolo reaffirmed that ―the MAU trend ha[d] already turned around,‖ remarked, ―we‘re in a great
3
place there,‖ and attributed the Q1 turnaround to ―a combination of seasonality or return to organic
4
growth and the set of product initiatives [Twitter] created to drive growth.‖ Compl. ¶ 88; ECF No.
5
92-7 at 8. The takeaway for analysts was ―reduced MAU growth targets but still [] very strong
6
growth for the remainder of 2015.‖ Compl. ¶ 105 (Morgan Stanley analyst explained ―[w]e are
7
reducing our full year 2015 core [Twitter] MAU expectations by 12mn (now modeling 16mn, 9mn
8
and 6mn net additions the next 3 qtrs)‖).
Defendants also painted a positive picture of user engagement, although investors were
9
given limited information. When one analyst asked directly whether user engagement was
11
United States District Court
Northern District of California
10
―improving or declining,‖ Noto gave the following response:
12
[A]s we think about engagement, there are a number of different ways that we
measure engagement – there‘s no one perfect way . . . . Additionally, on the
consumer side, many companies use DAU to MAU. And while that is a long-term
goal of ours, to become a daily product, today we have great variance in DAU to
MAU across geographies. In our more mature markets, we have very high DAU to
MAU, 50% plus. In the emerging markets, we have very low DAU to MAU, at
20% range. They all migrate up to a higher rate over time. And so as we get to a
point where we have a metric that‘s going to really reflect what we‘re trying to do,
we‘ll share that with you. But, at this point, there‘s a number of them that we look
at it, and no one metric to share.
13
14
15
16
17
18
19
20
21
22
23
Id. ¶ 84. To compare, at Analyst Day, Twitter reported a 48% DAU to MAU ratio in its top 20
markets. Id. ¶ 85(j). Then, despite describing Timeline views as a ―measurement that doesn‘t
reflect the initiatives that we‘re doing,‖ Noto touted that metric as indicative of improving user
engagement, explaining that ―timelines views per MAU totaled 631, up 3% year-over-year and
better than our outlook for timeline view per MAU to be flat versus Q4 2013.‖ Compl. ¶ 86; ECF
No. 92-7 at 6.
The message analysts took away from the 4Q 2014 results was clear: user engagement had
24
25
improved. Id. ¶¶ 52, 85(d) n.22, 86 (―[S]lower MAU growth is more than offset by improvements
26
9
27
28
To state the obvious, if MAU had grown by 47 million over the entire year but only 4 million of
those users joined in the last quarter, a slowdown of MAU growth had occurred at the end of
2014. A 4 million user increase is also notable when juxtaposed with Twitter‘s ambitious MAU
growth goals, discussed above.
10
1
in engagement and pricing.‖) (―MAU Growth about to Pick Up as Engagement Improves –
2
Raising PT to $65 . . . engagement rate growth accelerated.‖) (―Engagement and monetization
3
exceeded expectations.‖).
Twitter offered no new hard user engagement data in 2014 Form 10-K, filed on March 2,
4
5
2015. Instead, Twitter stated generally that its ―future revenue growth will depend on . . . our
6
ability to . . . increase user engagement.‖ Id. ¶ 44. Notably, Twitter‘s ―note regarding key
7
metrics‖ contained no reference to DAU:
8
We review a number of metrics, including monthly active users, or MAUs, timeline
views, timeline views per MAU and advertising revenue per timeline view, to
evaluate our business, measure our performance, identify trends affecting our
business, formulate business plans and make strategic decisions.
9
10
United States District Court
Northern District of California
11
12
13
14
15
16
17
Id. ¶ 91. This despite referring multiple times in the Form 10-K to the importance of user
engagement. E.g., id. ¶ 85(b) (―We believe that our future revenue growth will depend on, among
other factors, our ability to attract new users, increase user engagement and ad engagement.‖).10
Additionally, at an investor conference on March 3, Noto described user retention and preventing
disengaged users from dropping off the platform as Twitter‘s ―number one opportunity . . .
something we remain very focused on, it‘s a number we all look at everyday as an operating
committee.‖ Id. ¶ 28 n.8; ECF No. 106-9 at 9.
The lack of specific DAU data in Twitter‘s February and March filings caught the attention
18
19
20
of analysts and officials from the Securities and Exchange Commission (―SEC‖). An analyst from
RBC Capital Markets explained the problem:
If timeline views are out then analysts naturally want a replacement. Less
disclosure is always a bad thing especially for something that‘s important like
engagement . . . . I am totally open to the argument that [timeline views is] not a
useful metric. The question is: What‘s that new metric?
21
22
23
Id. ¶ 42; see also ¶ 49 (―Usage is the Key Underlying Growth Driver, But Lacking Real
24
Metrics . . . . MAUs don‘t tell the whole story . . . . Given that, we would also like to see Twitter
25
report DAUs (Daily Active Users) on a regular basis at some point, since that seems to be the
26
27
28
10
Twitter similarly emphasized user engagement in its Q1 2015 Form 10-Q, dated May 11, 2015.
Id. ¶ 104(a).
11
1
closest indication of usage frequency that could be externally reported in the near term.‖
2
Rosenblatt Securities (June 17, 2015)). Similarly, in April 2015, the SEC reminded Twitter of the
3
need to disclose ―metric(s) . . . to explain trends in user engagement,‖ citing its disclosure rules
4
that required Twitter to ―identify and discuss key performance indicators . . . that . . . management
5
uses to manage the business and that would be material to investors.‖ Id. ¶ 42.11
Despite receiving the SEC‘s letter, on Twitter‘s Q1 2015 Earnings Call on April 28, 2015,
6
7
Noto continued to tell investors that the Company could not provide ―visibility‖ on user
8
engagement. In response to a question about how analysts ―should be thinking about monitoring
9
engagement,‖ Noto said:
10
In terms of engagement metrics, there‘s a lot of different metrics that we look at
internally. There‘s not one metric for engagement. And so I can give you a sense of
some of them and quite frankly, we would like to be able to give you more
visibility on this, but there‘s just a number of different measurements. DAU is one
measurement of engagement.
United States District Court
Northern District of California
11
12
13
Id. ¶¶ 42, 103. Noto gave this explanation for Twitter‘s decision not to report specific DAU
14
numbers: ―It‘s a measurement that is dependent by market and you can have mixed shifts, so it
15
could be a little bit misleading.‖ Id. The only concrete information that Noto provided about
16
DAU on the April 28 call was that ―DAU to MAU ratios in the quarter were similar to what they
17
were‖ at Analyst Day. Id.
Noto attributed strong user engagement in part to Twitter‘s new product initiatives. On the
18
19
same call, he said ―[w]e‘re seeing perhaps the most exciting results [from a new initiative called
20
Recap] . . . These tweets are not only seeing higher engagement, they‘re bringing people back to
21
Twitter more frequently. Importantly, the machine learning work we‘re doing for Recap is
22
helping us to make these algorithms better and driving continuous improvements in engagement.‖
23
Id. ¶ 103.12 More generally, Noto and Cosolo reiterated that one of Twitter‘s three key objectives
24
11
25
26
27
28
The letter referred Twitter to SEC Release 33-8350 Section III.B. That release advises that,
―when preparing [the management discussion and analysis of an annual report], companies should
consider whether disclosure of all key variables and other factors that management uses to manage
the business would be material to investors, and therefore required.‖
12
Noto ―confirmed the success of the same new product initiative at an investor conference on
May 20, 2015, ―Dick talked about While YouWere Away on the earnings call and that we have
seen an improvement in engagement . . . .‖ Id. ¶ 103 n. 43.
12
1
was ―strengthen[ing] Twitter‘s core,‖ which ―represents our focus to increase the number of
2
logged-in users and their daily twitter use by increasing engagement and improving retention.‖ Id.
3
¶ 127.
Despite Twitter‘s positive statements about MAU and user engagement, the Company
4
struggled during the Class Period with stalled MAU growth numbers and ―increasingly adverse
6
trends in user engagement.‖ Id. ¶ 51. Multiple CWs confirm the Company‘s MAU and DAU
7
problems. According to CW-2, ―everyone at the Company with whom he spoke understood the
8
[MAU] growth was flat‖ between the summer of 2014 and the summer of 2015. Id. ¶ 68.
9
Similarly, CW-4 stated that at bi-weekly companywide meetings known as ―Tea Time,‖ ―Twitter
10
executives reported that MAU and DAU trends were flattening out.‖ Id. ¶ 70. CW-5 ―observed
11
United States District Court
Northern District of California
5
that DAU trends were declining during the Class Period‖ and that ―user base was flat or declining
12
leading up to and continuing through the Class Period.‖ Id. ¶ 71; see also id. ¶ 73 (CW-7) (―[B]y
13
late 2014, DAU growth was also mostly flat.‖); (id. ¶ 74) (CW-8) (―[I]n early 2015, growth was
14
‗really dying down.‘‖); (id. ¶ 77) (CW-11) (―[I]t was ‗inescapable‘ in late 2014 and early 2015
15
that there was ‗a lack of engagement‘ by users with the platform.‖). Many of the CWs reported
16
that ―it was apparent internally‖ during the Class Period that ―the MAU growth trend was flat.‖
17
Id. ¶ 98. For example, CW-2 explained that ―everyone internally‖ understood the MAU growth
18
was flat during CW-2‘s year-long tenure at Twitter (mid-2014 through mid-2015). Id. Indeed, the
19
Vanity Fair article13 describes how leadership stopped displaying an MAU chart with actual and
20
projected MAU at Twitter‘s Tea Time meetings because the ―gap between reality and hope [had
21
grown] so extreme.‖ Id. ¶ 78.
22
According to the CWs, Twitter concealed the fact that its newer users ―were less engaged
23
than more seasoned users and were more susceptible‖ to dropping off of the Twitter platform. In
24
other words, Twitter hid that new users were unlikely to become DAUs, risking future MAU
25
growth. CW-1, who was employed as a senior manager of the Growth and Engineering teams
26
27
28
13
The article, entitled ―Twitter is Betting Everything on Jack Dorsey. Will It Work?‖ and written
by Nick Bilton, was published in the summer of 2016. Id. ¶ 78.
13
1
until late 2015, explained that ―the MAU growth achieved during 2014 mostly involved ‗bringing
2
back low-quality MAU‘ and that low-quality MAUs were less likely to become DAUs.‖ Id. ¶ 67.
3
―[B]ecause Twitter concealed adverse trends in user engagement, investors could not gauge the
4
quality of MAU growth and were misled as to MAU growth trends.‖ Id. ¶ 58.
Twitter used various strategies to inflate its MAU numbers. For example, ―the Company
5
6
could increase its MAU in a given month by prompting users whose accounts had become inactive
7
to log in.‖ Id. ¶ 57; see also id. ¶ 69 (Vanity Fair article corroborating this technique). These
8
―zombie users‖ ―were more likely to drop off the platform‖ than ―organic growth MAUs‖ who
9
signed on to Twitter unprompted. Id. CW-3 agreed ―that ‗zombie users‘ and robot users
contributed to Twitter‘s overall MAU metric.‖ Id. ¶ 69. CW-6 noted that the existence of ―fake
11
United States District Court
Northern District of California
10
accounts contributed greatly to the number of ‗new‘ users and active users,‖ thus ―falsely
12
inflat[ing] Twitter‘s overall metrics.‖ Id. ¶ 72. ―When CW-6 reported the issue of the fake
13
accounts and presented the business case to CW-6‘s manager, he was told to ‗do your job and be
14
quiet.‘‖ Id. Additionally, ―automated third party users,‖ who ―did not actually log in to Twitter
15
and use the platform [] were still counted as MAUs as the result of a third party application
16
automatically pinging Twitter‘s servers for updates.‖ Id. ¶ 87(c).14 Investors reported on these
17
―auto-pulling‖ users, with one calling them ―a point of interest . . . as it means a large amount of
18
the . . . active monthly users Twitter reported . . . are not real consumers that have the potential to
19
view and interact with their advertising and organic content.‖ Id. ¶ 89(e).
Low DAU and MAU growth forced Twitter to alter its advertising strategy. In late 2014,
20
21
according to CW-7, management increased the limits on the number of ads a user could see in a
22
given time frame. Id. ¶ 73. ―[I]n order to improve the advertisement metrics and the number of
23
advertisements sent to users, the ‗Ad Load‘15 had to be increased to compensate for the lack of
24
25
14
26
In Q4 2014, Apple issued an iPhone software update that changed this automated activity and
cost Twitter 3 million users who had previously contributed to the Company‘s DAU and MAU
numbers. Id. ¶ 87(d).
27
15
28
―Ad load‖ refers to ―the amount of advertisements that would appear on the [Twitter] platform.‖
Id. at 6 n.5.
14
1
user (MAU) growth and user engagement (DAU) growth.‖ Id. ¶ 73. Twitter then pointed analysts
2
and SEC staff to ad engagements as a metric that is ―helpful to investors to understand‖ and
3
―monitor trends in user engagement,‖ id., despite the fact that ―ad engagements (a monetization
4
metric)16 was moving in the opposite direction from the trend in user engagement during the Class
5
Period.‖ Id. ¶ 85(i). For example, Twitter reported that ad engagements grew by 70% year over
6
year in Q4 2014, and 32% year over year in Q1 2015. Id. ¶ 85(i). Twitter also responded to the
7
SEC‘s request for a disclosure of a reliable user engagement metric with the following comment:
8
―The Company respectfully advises the Staff that it has included two metrics, changes in ad
9
engagements and changes in cost per ad engagement.‖ Id. ¶ 110(c). But according to CW-11,
linking user engagement with ad engagement was misleading because ―there is no direct
11
United States District Court
Northern District of California
10
correlation between advertising engagement and MAU or DAU.‖ Id. In other words, ad
12
engagements was a poor measure of engagement because, as CW-7 explained, Ad Load could
13
always be ―increased to compensate for the lack of user (MAU) growth and user engagement
14
(DAU) growth.‖ Id. ¶ 73.
15
The weeks before the end of the Class Period saw the departure of several key Twitter
16
executives. First, Costolo ―was forced from his position as CEO on July 1, 2015.‖ Id. ¶ 129.
17
Both the CWs and investors attributed Costolo‘s resignation to the Company‘s problems with user
18
growth and engagement. Id. ¶¶ 130-31. Twitter‘s chief of corporate communications, Gabriel
19
Stricker, was the next to announce his departure on July 16, 2015. Id. ¶ 132. Other ―key
20
executives‖ followed. Id. ¶ 133 (Vice President of Corporate Development and Strategy, Rishi
21
Garg; Christian Oestlien, Vice President of Product Management; and Todd Jackson, a product
22
director). One analyst observed at the end of July 2015 that ―only 9 of the 13 @Twitter execs that
23
presented at their Nov 2014 analyst day are left.‖ Id. ¶ 133.
24
C.
End of the Class Period
25
On July 28, 2015, ―Twitter hosted a call with analysts and investors to discuss the
26
16
27
28
Defendants had previously referred to ad engagements as a monetization metric. Id. ¶ 115. For
example, on the April 28 Earnings Call, the Company gave this report: ―Turning now to
monetization metrics, . . . year-over-year ad revenue growth. . . was driven by both an increase in
ad engagements and [cost per ad engagement].‖ Id.
15
1
Company‘s second-quarter results.‖ Id. ¶ 60. In contrast to prior, upbeat forecasts, Noto told
2
investors: ―[W]e do not expect to see sustained meaningful growth in MAUs . . . [for] a
3
considerable period of time.‖ Id. Generally, Twitter reported that its ―growth rate in users is
4
slowing quite dramatically‖ and that investors should not ―expect a change in our growth rate . . .
5
for a while.‖ Id. More specifically, ―Noto reported that . . . user engagement had significantly
6
declined since Analyst Day, with the DAU to MAU ratio falling from 48% to 44%.‖ Id. ¶ 61. In
7
the same vein, ―Noto acknowledged that new product initiatives ‗ha[d] not yet had a meaningful
8
impact on growing our audience [i.e.,MAU] or participation [i.e., user engagement].‘‖ Id. ¶ 62.
9
Twitter also admitted that the lack of user engagement could negatively impact advertising
revenue: ―[I]f we do not grow audience, drive increased engagement or begin to monetize other
11
United States District Court
Northern District of California
10
areas such as Logged Out, it is possible that on some days our revenue could be impacted by
12
limited availability for specific ad types.‖ Id. ¶ 87 (g). In other words, because Twitter only sent
13
ads to a user‘s timeline when the user logged in to the platform, low user engagement was
14
constraining Twitter‘s ability to generate ad revenue. Id.
After the earnings call, Twitter‘s share price dropped nearly 15%. Id. ¶ 64.17 According to
15
16
the Vanity Fair article, the 2Q 2015 call represented a conscious decision to ―come clean‖ about
17
lagging user growth and engagement. Id. ¶ 78.18
18
D.
Post-Class Period Activity
19
After the July 28 revelations, Twitter no longer wavered on the centrality of DAU to the
20
Company‘s success. During the Q4 2015 Earnings Call, for example, Twitter explained that ―[t]he
21
one engagement metric that we look at holistically is daily active users.‖ Id. ¶ 85(b). Moreover,
22
Twitter reminded investors that the importance of user engagement only grows if MAU growth
23
decelerates. Id. ¶ 85(d) (Noto explains at September 16, 2015 conference that ―as your MAU
24
17
25
This Order does not include Plaintiff‘s factual allegations related to loss causation or the
presumption of reliance because Twitter does not address either issue in its motion to dismiss.
26
18
27
28
According to the Vanity Fair article, ―[i]n July 2015, Gabriel Stricker, the Company‘s director
of communications, told Dorsey, Noto, and other top managers: ‗We have zero credibility with
Wall Street right now . . . . We have to come clean about the company‘s stagnant growth
numbers.‘‖ Id. ¶ 89(c).
16
1
growth slows, engagement becomes a much bigger factor‖). Relatedly, Twitter finally disclosed
2
specific DAU numbers, illustrating the adverse user engagement trends that had existed
3
throughout the Class Period. ―On October 27, 2016, nearly 18 months after the Q1 2015 earnings
4
call, Defendants disclosed DAU growth for Q1 2016 versus Q1 2015.‖ Id. ¶ 102(a). The Q1 2015
5
DAU/MAU ratio was 43.8%, down over 4% from the 48% ratio reported in November 2014 at
6
Analyst Day. Id. Defendants also acknowledged that Twitter‘s new product initiatives had not
7
produced meaningful user engagement returns going as far back as September 2014. In
8
September 2015, the Company gave this report about ―the last year+‖:
9
Some of [the product initiatives] have statistically shown positive results, but they
haven‘t been impactful to the numbers. And the reality is, is for us to reach that
next cohort of users, we have to appeal to their needs, and we have to make the
product simple and easy enough. That requires a fundamental product change.
10
United States District Court
Northern District of California
11
12
Id. ¶ 89(g).
Qualifying its MAU growth, Twitter explained that its newer users were ―lower quality‖
13
14
15
16
17
18
than existing users, and therefore less likely to become DAUs. For example, Noto told investors
at a September 2015 investor conference that ―the more recent MAUs that we‘ve acquired. . . .
They‘re lower quality or lower conversion to DAU. And [we] just wanted to be transparent about
that.‖ Id. ¶ 87(h). Similarly in December 2015 Noto explained that ―the newer MAUs that we are
acquiring were not as engaged as our existing MAUs.‖ Id.
Analysts quickly reported on Twitter‘s adverse user engagement and DAU numbers. Two
19
20
21
22
23
24
25
26
27
28
weeks after the close of the Class Period, the Susquehanna Financial Group published a chart
showing a decline in the DAU/MAU ratio from an average of 48% during the first three quarters
of 2014 to 44% during the first quarter of 2015. Id. ¶ 102(c). Forbes noted that ―engagement []
has been declining since last fall,‖ and the Wall Street Journal reported that that user engagement
had ―gone south‖ since Analyst Day. Id. ¶ 87(b).
III.
LEGAL STANDARD
A.
Dual Pleading Requirements
―In a typical § 10(b) private action a plaintiff must prove (1) a material misrepresentation
or omission by the defendant; (2) scienter; (3) a connection between the misrepresentation or
17
1
omission and the purchase or sale of a security; (4) reliance upon the misrepresentation or
2
omission; (5) economic loss; and (6) loss causation.‖ Stoneridge Inv. Partners, LLC v. Sci.-
3
Atlanta, 552 U.S. 148, 157 (2008).
On a motion to dismiss, the Court accepts the material facts alleged in the complaint,
4
together with reasonable inferences to be drawn from those facts, as true. Navarro v. Block, 250
6
F.3d 729, 732 (9th Cir. 2001). However, ―the tenet that a court must accept a complaint‘s
7
allegations as true is inapplicable to threadbare recitals of a cause of action‘s elements, supported
8
by mere conclusory statements.‖ Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). Moreover, while a
9
plaintiff generally need only plead ―enough facts to state a claim to relief that is plausible on its
10
face‖ to survive a motion to dismiss, Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 570 (2007),
11
United States District Court
Northern District of California
5
―[s]ecurities fraud class actions must meet the higher, exacting pleading standards of Federal Rule
12
of Civil Procedure 9(b) and the Private Securities Litigation Reform Act (‗PSLRA‘).‖ Oregon
13
Pub. Employees Ret. Fund v. Apollo Grp. Inc., 774 F.3d 598, 604 (9th Cir. 2014).
Under the PSLRA and Rule 9(b), a complaint must ―state with particularity facts giving
14
15
rise to a strong inference that the defendant acted with the required state of mind‖ with respect to
16
each alleged false statement or omission, and a party must ―state with particularity the
17
circumstances constituting fraud or mistake.‖ 15 U.S.C. § 78u–4(b)(2)(A); Fed. R. Civ. P. 9(b);
18
see also Oregon Pub. Employees Ret. Fund, 774 F.3d at 605. ―In order to show a strong inference
19
of deliberate recklessness, plaintiffs must state facts that come closer to demonstrating intent, as
20
opposed to mere motive and opportunity.‖ In re Silicon Graphics Inc. Sec. Litig., 183 F.3d 970,
21
974 (9th Cir. 1999), abrograted on other grounds by, S. Ferry LP, No. 2 v. Killinger, 542 F.3d 776,
22
784 (9th Cir. 2008). If the complaint does not satisfy the PSLRA‘s pleading requirements, the
23
Court must grant a motion to dismiss the complaint. 15 U.S.C. § 78u–4(b)(3)(A).
Falsity and Materiality19
24
B.
25
The PSLRA provides that ―the complaint shall specify each statement alleged to have been
26
misleading, the reason or reasons why the statement is misleading, and, if an allegation regarding
27
28
19
Defendants do not challenge the materiality of their statements in the motion to dismiss.
18
1
the statement or omission is made on information and belief, the complaint shall state with
2
particularity all facts on which that belief is formed.‖ 15 U.S.C. § 78u–4(b)(1)(B). For statements
3
to be actionable under the PSLRA, they must be both false or misleading and material.
4
A statement or omission is misleading under the PSLRA and Section 10(b) of the
5
Exchange Act ―if it would give a reasonable investor the impression of a state of affairs that
6
differs in a material way from the one that actually exists.‖ Berson v. Applied Signal Tech., Inc.,
7
527 F.3d 982, 985 (9th Cir. 2008). ―[O]nly if ‗reasonable minds‘ could not disagree that the
8
challenged statements were not misleading should the district court dismiss under 12(b)(6).‖
9
Warshaw v. Xoma Corp., 74 F.3d 955, 959 (9th Cir. 1996).
10
A false or misleading statement or omission is material if there is a ―substantial likelihood
United States District Court
Northern District of California
11
that the disclosure of the omitted fact would have been viewed by the reasonable investor as
12
having significantly altered the ‗total mix‘ of information made available.‖ TSC Indus., Inc. v.
13
Northway, Inc., 426 U.S. 438, 449 (1976). ―To plead materiality, the complaint‘s allegations must
14
‗suffice to raise a reasonable expectation that discovery will reveal evidence satisfying the
15
materiality requirement, and to allow the court to draw the reasonable inference that the defendant
16
is liable.‘‖ Reese v. Malone, 747 F.3d 557, 568 (9th Cir. 2014) (quoting Matrixx Initiatives, Inc.
17
v. Siracusano, 563 U.S. 27 (2011)). ―‗Although determining materiality in securities fraud cases
18
should ordinarily be left to the trier of fact, conclusory allegations of law and unwarranted
19
inferences are insufficient to defeat a motion to dismiss for failure to state a claim.‘‖ Id. (quoting
20
In re Cutera Sec. Litig., 610 F.3d 1103, 1108 (9th Cir. 2010)). While statements made before or
21
after the class period are not themselves actionable, they may be relevant in that they shed light on
22
the ―truth or falsity of Class Period statements.‖ In re Invision Techs., Inc. Sec. Litig., No. C04-
23
03181 MJJ, 2006 WL 538752, at *2 (N.D. Cal. Jan. 24, 2006).
24
C.
Scienter
25
The required state of mind under the PSLRA is a ―mental state embracing intent to
26
deceive, manipulate, or defraud.‖ Ernst & Ernst v. Hochfelder, 425 U.S. 185, 193–94 n.12 (1976).
27
In order to adequately establish scienter, the complaint must ―state with particularity facts giving
28
rise to a strong inference that the defendant acted with the required state of mind.‖ 15 U.S.C. §
19
1
78u–4(b)(2)(A). The ―strong inference‖ required by the PSLRA ―must be more than merely
2
‗reasonable‘ or ‗permissible‘—it must be cogent and compelling, thus strong in light of other
3
explanations.‖ Tellabs, Inc. v. Makor Issues & Rights, Ltd., 551 U.S. 308, 324 (2007). ―A court
4
must compare the malicious and innocent references cognizable from the facts pled in the
5
complaint, and only allow the complaint to survive a motion to dismiss if the malicious inference
6
is at least as compelling as any opposing innocent inference.‖ Zucco Partners, LLC v. Digimarc
7
Corp., 552 F.3d 981, 991 (9th Cir. 2009). In evaluating whether a complaint satisfies the ―strong
8
inference‖ requirement, courts must consider the allegations and other relevant material
9
holistically, not ―scrutinized in isolation.‖ In re VeriFone Holdings, 704 F.3d 694, 701 (9th Cir.
10
2012).
Deliberate or conscious recklessness constitutes intentional conduct sufficient to satisfy the
United States District Court
Northern District of California
11
12
scienter requirement. ―An actor is deliberately reckless if he had reasonable grounds to believe
13
material facts existed that were misstated or omitted, but nonetheless failed to obtain and disclose
14
such facts although he could have done so without extraordinary effort.‖ Reese, 747 F.3d at 569
15
(quoting In re Oracle Corp. Sec. Litig., 627 F.3d 376, 390 (9th Cir. 2010) (internal alterations
16
omitted)). ―[T]he ultimate question is whether the defendant knew his or her statements were false,
17
or was consciously reckless as to their truth or falsity.‖ Gebhart v. SEC, 595 F.3d 1034, 1042 (9th
18
Cir. 2010). ―Facts showing mere recklessness or a motive to commit fraud and opportunity to do
19
so provide some reasonable inference of intent, but are not independently sufficient.‖ Reese, 747
20
F.3d at 569 (quoting In re Silicon, 183 F.3d at 974).
21
IV.
22
DISCUSSION
Plaintiff‘s Complaint contains dozens of allegedly false or misleading Class Period
23
statements. The challenged statements fall into three broad categories: 1) claims based on the
24
omission of DAU metrics, 2) claims based on affirmative statements about user engagement, and
25
3) claims based on affirmative statements about MAU. The Court addresses each category of
26
statement, then discusses scienter separately.
27
A.
Claim Based on Omission of DAU Metrics
28
Plaintiff‘s strongest argument is that Twitter misled investors by failing to disclose DAU
20
1
metrics during the class period. Twitter responds by arguing that it had no duty to disclose DAU
2
metrics. ECF No. 91-1 at 15. Twitter is correct that, as a general matter, ―Section 10(b) and Rule
3
10b–5(b) do not create an affirmative duty to disclose any and all material information.‖ In re
4
Mellanox Techs. Ltd. Sec. Litig., No. 13-CV-04909-JST, 2014 WL 12650991, at *17 (N.D. Cal.
5
Mar. 31, 2014) (citing Matrixx Initiatives, Inc., 131 S. Ct. at 1321). It is well settled, however,
6
that a duty to disclose does arise where an omission is misleading because it ―affirmatively
7
create[s] an impression of a state of affairs that differs in a material way from the one that actually
8
exists.‖ Brody v. Transitional Hosps. Corp., 280 F.3d 997, 1006 (9th Cir. 2002). In other words,
9
while Twitter had no duty to disclose in the abstract, the Company‘s omission of DAU data is
10
actionable if it was misleading.20
Plaintiff argues that omission of DAU was misleading for three main reasons. First, the
United States District Court
Northern District of California
11
12
Complaint alleges that non-disclosure of DAU was misleading because ―Twitter management
13
relied on DAU as its primary user engagement metric during the Class Period.‖ Id. ¶ 104(d). As
14
an initial matter, Plaintiff has plausibly alleged that DAU was in fact Twitter‘s main user
15
engagement metric. At Analyst Day in November 2014, Twitter unveiled an ―operational goal of
16
building the world‘s largest daily audience.‖ Id. ¶ 31. Several months later, in response to an
17
analyst question whether ―people [should] still focus on monthly active users,‖ Noto explained
18
that Twitter‘s ―number-one priority is to strengthen the core and make Twitter an increasingly
19
daily use case for the people who come to Twitter.‖ Id. ¶ 33. Twitter‘s SEC filings during the
20
Class Period continuously referred to the importance of user engagement, e.g., id. ¶ 85(c) (―We
21
believe that our future revenue growth will depend on, among other factors, our ability to attract
22
new users, increase user engagement and ad engagement.‖). Finally, after the Class Period ended,
23
Noto stated clearly that ―[t]he one engagement metric that we look at holistically is daily active
24
20
25
26
27
28
Plaintiff also argues that SEC Regulation S-K required disclosure of key operating metrics.
Compl. ¶¶ 91-93 (referencing 17 C.F.R. § 229 and SEC Release 33-8350). The Court agrees with
Twitter, however, that these regulations do not create an affirmative duty to disclose. See In re
NVIDIA Corp. Sec. Litig., 768 F.3d 1046, 1056 (9th Cir. 2014) (―In sum, we hold that Item 303
does not create a duty to disclose for purposes of Section 10(b) and Rule 10b–5. Such a duty to
disclose must be separately shown according to the principles set forth by the Supreme Court in
Basic and Matrixx Initiatives.‖).
21
1
users,‖ and ―as it relates to engagement . . . , [t]he one that is probably the most important is daily
2
active users.‖ Id. ¶ 85(b). The CWs‘ testimony and analysts‘ reporting confirms the centrality of
3
DAU. Id. ¶ 39 (analyst reports that ―TWTR set some ambitious goals as it aims to have the largest
4
DAU in [the] world with over $14 billion of annual revenue‖); id. ¶ 67 (―CW-1 reported that DAU
5
was the primary engagement metric that Twitter tracked internally after Timeline Views were no
6
longer being reported.‖). In sum, taking the facts alleged in the Complaint as true, starting months
7
before the Class Period, DAU was Twitter‘s primary user engagement metric.
8
But just because DAU was Twitter‘s key engagement metric, it does not necessarily follow
9
that its omission ―affirmatively create[s] an impression of a state of affairs that differs in a material
way from the one that actually exists.‖ Brody, 280 F.3d 997, 1006. To make this showing,
11
United States District Court
Northern District of California
10
Plaintiff points the Court to Defendants‘ statements that Twitter tracks user engagement in more
12
than one way. For example, on February 5, 2014, when one analyst asked directly whether user
13
engagement was ―improving or declining,‖ Noto gave the following response:
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
[A]s we think about engagement, there are a number of different ways that we
measure engagement – there‘s no one perfect way. . . . Additionally, on the
consumer side, many companies use DAU to MAU. And while that is a long-term
goal of ours, to become a daily product, today we have great variance in DAU to
MAU across geographies. In our more mature markets, we have very high DAU to
MAU, 50% plus. In the emerging markets, we have very low DAU to MAU, at
20% range. They all migrate up to a higher rate over time. And so as we get to a
point where we have a metric that‘s going to really reflect what we‘re trying to do,
we‘ll share that with you. But, at this point, there‘s a number of them that we look
at it, and no one metric to share.
Id. ¶ 84. Similarly, on the April 28 Earnings Call, Noto told investors the following:
In terms of engagement metrics, there‘s a lot of different metrics that we look at
internally. There‘s not one metric for engagement. And so I can give you a sense of
some of them and quite frankly, we would like to be able to give you more
visibility on this, but there‘s just a number of different measurements. DAU is one
measurement of engagement.
Id. ¶¶ 42, 103. The Court does not see how these statements are actionable. The difference
between ―DAU is one measure of engagement‖ and the hypothetical ―DAU is the primary measure
of engagement‖ is too small to render the first statement misleading. Nor is it inconsistent to
report that there are ―a lot of different metrics‖ for user engagement just because one of those
statements is more closely tracked than the others. In sum, these statements do not render
22
1
2
Twitter‘s omission of specific DAU data misleading.
Second, Plaintiff argues that omission of the DAU metric was misleading when set against
3
Twitter‘s statements regarding MAU growth. Specifically, Plaintiff claims that ―without DAU,
4
investors were led to believe that Defendants‘ outsized MAU projections were viable, MAU
5
growth was high quality (i.e., new users were just as engaged as existing users), and new product
6
features designed to increase growth were working.‖ ECF No. 94 at 14. Twitter responds that it
7
has ―no independent legal duty to provide ‗context‘‖ for its MAU numbers, and that to hold
8
otherwise would ―run[] afoul of the principle that securities laws ‗do[] not impose a duty of
9
completeness.‘‖ ECF No. 91-1 at 18 (quoting City of Royal Oak Ret. Sys. v. Juniper Networks,
880 F. Supp. 2d 1045, 1066 (N.D. Cal. 2012)). Again, although this Court acknowledges that
11
United States District Court
Northern District of California
10
there is no duty of completeness, a plaintiff can state a 10(b) claim based on a failure to provide
12
―context‖ where that failure ―affirmatively create[s] an impression of a state of affairs that differs
13
in a material way from the one that actually exists.‖ Brody, 280 F.3d at 1006. Indeed, Twitter
14
acknowledges the viability of this theory in its motion. ECF No. 91-1 at 18 ((―Although a duty to
15
disclose could attach if MAU data was misleading absent DAU disclosure . . . .‖).
16
The picture as to this theory of relief is different. To succeed on this theory, Plaintiff must
17
allege that Twitter reported positive MAU growth, that Twitter was simultaneously experiencing
18
adverse DAU trends, and that those DAU trends made MAU growth implausible. Plaintiff has
19
sufficiently pleaded all three steps. Plaintiff alleged that Twitter told investors that it was on track
20
to produce the kind of MAU growth Company officials described at Analyst Day. During the
21
February 5 Earning Call, Costolo acknowledged lower than expected MAU growth at the end of
22
2014, but described signs of a rebound in early 2015:
23
24
25
26
27
28
We ended the quarter with 288 million monthly active users. We added 4 million
users this quarter and 47 million across 2014. There are quarter-specific factors that
impacted our net adds in Q4, which includes seasonality and a couple issues related
to the launch of iOS 8. We'll discuss that in more detail later in this call.
Importantly, I want to highlight that the user numbers we saw in January of
this year indicate that our MAU trend has already turned around and our Q1
trend is likely to be back in the range of absolute net adds that we saw during the
first three quarters of 2014.
Id. ¶ 88; ECF No. 92-7 at 3 (emphasis added). When asked about the claims of ―acceleration‖ in
23
1
MAU growth, Costolo said again that the ―MAU trend ha[d] already turned around,‖ remarked,
2
―we‘re in a great place there,‖ and attributed the Q1 turnaround to ―a combination of seasonality
3
or return to organic growth and the set of product initiatives [Twitter] created to drive growth.‖
4
Compl. ¶ 88; ECF No. 92-7 at 8. These statements convey a message of that MAU was trening
5
positively, and they were interpreted by analysts as such. Id. ¶ 105 (reporting ―reduced MAU
6
growth targets but still [] very strong growth for the remainder of 2015); id. ¶ 86 (―[S]lower MAU
7
growth is more than offset by improvements in engagement and pricing.‖) (―MAU Growth about
8
to Pick Up as Engagement Improves – Raising PT to $65 . . . engagement rate growth
9
accelerated.‖) (―Engagement and monetization exceeded expectations.‖).
Simultaneously, Twitter was experiencing flat or declining DAU trends and other
11
United States District Court
Northern District of California
10
problems with user engagement. As Defendants admitted after the end of the Class Period, ―user
12
engagement had significantly declined since Analyst Day, with the DAU to MAU ratio falling
13
from 48% to 44%.‖ Id. ¶ 61. In their motion and on reply, Defendants challenge Plaintiff‘s
14
calculation of these figures and the inferences to be drawn from them. For example, Defendants
15
note that ―the analyst report [from which Plaintiff obtained the 44% figure] actually compares a
16
ratio for the first three quarters of 2014 with a ratio for the second quarter of 2015.‖ ECF No. 91-1
17
at 23. As a result, Defendants argue, Plaintiff cannot allege precisely when the ratio declined.
18
ECF No. 104 at 8. But at the motion to dismiss phase, Plaintiff‘s numbers give rise to the
19
inference that the DAU to MAU ratio was declining during the Class Period. The fact that
20
Defendants‘ alleged obfuscation of their metrics, Compl. ¶ 102(a) n.41, makes it difficult to say
21
when within the Class Period the decline occurred does not defeat that inference. Defendants also
22
argue that a decline in the DAU to MAU ratio does not necessarily mean the total number of
23
DAUs also declined. ECF No. 104 at 9. But Plaintiff did allege a decline in overall DAU from
24
Analyst Day to Q1 2015. Compl. ¶ 102(a) (from 109 million to 105.8 million DAU). Nor do
25
Defendants‘ mathematical arguments persuade the Court that it is ―[i]t is entirely reasonable that
26
DAUs increased but at a slower rate than the MAU increase.‖ ECF No. 104 at 9. For example,
27
Twitter compares DAU to MAU ratios for its top 20 markets with MAU numbers for the entire
28
company to extrapolate DAU growth. Id. at 9. Twitter criticized Plaintiff for exactly this type of
24
1
supposedly improper comparison. E.g., ECF No. 91 at 25 n.13. All this goes to show why these
2
debates over the accuracy of Plaintiff‘s numbers are not appropriate at the motion to dismiss
3
phase.
In addition to this allegation about the DAU to MAU ratio, in September 2015, Noto told
4
5
investors that ―the more recent MAUs that we‘ve acquired. . . . They‘re lower quality or lower
6
conversion to DAU. And [we] just wanted to be transparent about that.‖ Id. ¶ 87(h). This
7
statement and the allegations of a declining DAU to MAU ratio are consistent with the testimony
8
of the CWs, many of whom reported adverse DAU trends and problems with user engagement
9
during the Class Period. Id. ¶ 70. (CW-4 stated that at Tea Time meetings, ―Twitter executives
reported that MAU and DAU trends were flattening out.‖) (CW-5 ―observed that DAU trends
11
United States District Court
Northern District of California
10
were declining during the Class Period‖ and that ―user base was flat21 or declining leading up to
12
and continuing through the Class Period.‖).
Plaintiff claims that Twitter should have disclosed these adverse DAU and user
13
14
engagement trends because those trends would have allowed investors to understand that Twitter‘s
15
statements about MAU acceleration were unrealistic. Plaintiff has plausibly alleged, through
16
statements by Defendants, the CWs, and analysts, that MAU is unhelpful at best and misleading at
17
worst in the absence of companion DAU data. Twitter acknowledged the link between DAU and
18
MAU years prior to the class period in its 2013 Registration Statement, explaining that ―[t]o the
19
extent our user growth rate slows, our success will become increasingly dependent on our ability
20
to increase levels of user engagement.‖ Id. ¶ 24. The CWs agreed. Id. ¶ 75 (CW-9) (―MAU says
21
nothing about frequency‖ of use and therefore does not provide ―sufficient detail to measure an
22
advertising opportunity.‖); Id. ¶ 76 (CW-10) (―‘MAU was a terrible metric‘ and there is no way
23
you could ‗judge the health of the company using MAU alone.‘‖). Finally, as one tech analyst
24
explained, MAU metrics ―indicate very little about true engagement on a platform, because using
25
an app every 30 days isn‘t that much different from never using it at all. For social and
26
communication apps, the key is daily usage because it tells you how people are really engaging.‖
27
28
21
Plaintiffs allege that flat DAU was considered a negative trend. Id. ¶ 87(a).
25
1
Id. ¶ 121 (internal alterations omitted).
Putting these elements together, Plaintiff has plausibly alleged that, set against Defendants‘
2
claims that MAU growth had ―turned around‖ in early 2015, the omission of DAU was
4
misleading. As the Ninth Circuit has explained, ―‗once defendants cho[o]se to tout‘ positive
5
information to the market, ‗they [are] bound to do so in a manner that wouldn't mislead investors,‘
6
including disclosing adverse information that cuts against the positive information.‖ Schueneman
7
v. Arena Pharm., Inc., 840 F.3d 698, 705–06 (9th Cir. 2016) (quoting Berson v. Applied Signal
8
Tech., Inc., 527 F.3d 982, 987 (9th Cir. 2008)). In re LendingClub Securities Litigation, from this
9
district, is also instructive. No. C 16-02627 WHA, 2017 WL 2289186 (N.D. Cal. May 25, 2017).
10
There, defendant LendingClub ―operated an online peer-to-peer marketplace to match borrowers
11
United States District Court
Northern District of California
3
and investors for a variety of loans.‖ Id. at *1. The plaintiffs challenged, among other things, the
12
omission on LendingClub‘s IPO registration statement the existence of a related company called
13
Cirrix, which was formed for ―the sole purpose of purchasing loans from LendingClub.‖ Id. at *7.
14
The court agreed the omission was misleading, noting that ―investors would also have been
15
interested to know that millions of dollars of loans were not the result of organic matches in the
16
marketplace reflecting participants‘ trust in LendingClub‘s purportedly neutral platform, but were
17
inflated artificially through self-dealing disguised as real transactions.‖ Id. at *8. Likewise, here,
18
disclosure of the fact that DAU was flat or declining during the Class Period would have put
19
investors on alert that MAU was not really accelerating in 2015 and that the aggressive MAU
20
projections announced at Analyst Day were unlikely to materialize.22 In the absence of DAU data,
21
investors interpreted Defendants‘ statements as reassurances that the Company had experienced
22
and would continue to experience positive growth and engagement trends. Compl. ¶ 86
23
(―[S]lower MAU growth is more than offset by improvements in engagement and pricing.‖)
24
(―MAU Growth about to Pick Up as Engagement Improves – Raising PT to $65 . . . engagement
25
26
22
27
28
Relatedly, Plaintiff claims that the omission of DAU was misleading because it obscured the
fact that Twitter‘s MAU growth was not high quality, ―organic‖ growth. As discussed below,
however, these statements about the quality of MAU growth are not actionable.
26
1
rate growth accelerated.‖) (―Engagement and monetization exceeded expectations.‖).23 The
2
omission was therefore misleading.
Third, Plaintiff claims the omission of DAU was misleading when set against Twitter‘s
3
4
positive statements about user engagement trends. ECF No. 94 at 17. The Court first rejects
5
Defendants‘ response that there was no adverse user engagement trend in the first place. ECF No.
6
91-1 at 6. As discussed above, Plaintiff adequately alleged flat or declining DAU and other
7
problems with user engagement. Despite those negative trends, Defendants‘ Class Period
8
statements suggested user engagement was improving. During the February 5, 2015 Earnings
9
Call, for example, an analyst asked Noto directly whether ―engagement was improving or
10
declining.‖ Compl. ¶ 86. Noto gave the following response:
United States District Court
Northern District of California
11
In our more mature markets, we have very high DAU to MAU, 50% plus. In the
emerging markets, we have very low DAU to MAU, at 20% range. They all
migrate up to a higher rate over time. And so as we get to a point where we have a
metric that‘s going to really reflect what we‘re trying to do, we‘ll share that with
you. But, at this point, there‘s a number of them that we look at it, and no one
metric to share.
12
13
14
15
16
17
18
19
20
Id. ¶ 84. To compare, at Analyst Day, Twitter reported a 48% DAU to MAU ratio in its top 20
markets. Noto‘s comment therefore suggests an increase in user engagement.24 In response to the
same analyst question, Noto also touted improvements in Twitter‘s outdated Timeline Views
metric, explaining that ―timelines views per MAU totaled 631, up 3% year-over-year and better
than our outlook for timeline view per MAU to be flat versus Q4 2013.‖ Compl. ¶ 86; ECF No.
92-7 at 6.
The most plausible takeaway from these statements was that Twitter‘s user engagement
21
22
23
24
25
23
This case is therefore distinguishable from Bodri v. GoPro, Inc., No. 16-CV-00232-JST, 2017
WL 1732022, at *10 (N.D. Cal. May 1, 2017), because Plaintiff plausibly alleges that reasonable
investors did interpret Defendants‘ statements as predicting ―accelerating‖ user engagement and
growth.
24
26
27
28
Defendants take issue with this comparison, arguing that Noto was referring on February 5 to
Twitter‘s top five markets, not its top twenty markets. ECF No. 91-1 at 25. This amounts to a
dispute over the inferences to be drawn from Noto‘s two statements. On a motion to dismiss, the
Court draws reasonable inferences in favor of Plaintiff, and finds that Noto‘s February 5 statement
amounted to a claim that the Company‘s DAU numbers had improved.
27
1
was improving and would continue to do so.25 See id. ¶ 86 (analysts summarizing Defendants‘
2
Class Period statements as reporting accelerating user engagement). Yet Plaintiff plausibly alleged
3
that user engagement, and DAU in particular, was flat or declining leading up to and during the
4
Class Period. As with MAU, ―once defendants cho[o]se to tout positive information to the
5
market,‖ about user engagement, they were ―bound to do so in a manner that wouldn't mislead
6
investors, including disclosing adverse information that cuts against the positive information.‖
7
Schueneman, 840 F.3d 698 at 705–06. Here, that meant disclosing specific DAU numbers, which
8
would have revealed that Noto‘s February 5 statements suggesting positive DAU and user
9
engagement trends were misleading.
In sum, the Court denies Defendants‘ motion to dismiss Plaintiff‘s claim based on the
10
United States District Court
Northern District of California
11
omission of DAU metrics.
12
B.
13
Claim Based on Affirmative Statements about Positive User Engagement
Trends
14
Next, Plaintiff claims to have stated a plausible 10(b) claim based on Twitter‘s affirmative
15
16
statements about positive user engagement trends. Many of these statements are the same ones
17
Plaintiff points to above to support their DAU omission claim. For example, Plaintiff again refers
18
the Court to the Q4 2014 Earnings Call on February 5, 2015, when one investor asked whether
19
―engagement was improving or declining.‖ Id. ¶ 84. Noto responded by reporting a 3% increase
20
in the outdated Timeline Views metric and ―very high DAU to MAU, 50% plus‖ in Twitter‘s
21
―more mature markets.‖ Id. The Court already explained above why these statements were
22
misleading absent an accompanying disclosure of more specific DAU data. Plaintiff has not
23
plausibly alleged, however, that these statements were false in and of themselves. Just because
24
25
26
27
28
25
Twitter objects that this ―historical‖ statement says nothing about future growth. ECF No. 104
at 10. The Court disagrees. First, the comment was made in response to a question about whether
growth was ―improving.‖ Second, giving a DAU to MAU ratio that was allegedly higher than a
previous DAU to MAU ratio indicates growth, as does the statement that the ratios ―migrate up to
a higher rate over time.‖ In any event, Plaintiff also alleges that the omission of DAU was
misleading as to current user engagement, not just future trends.
28
1
Twitter had discontinued Timeline Views does not make a statistical statement about that metric
2
false. Similarly, Plaintiff has not stated that Noto‘s ―50% plus‖ comment was factually inaccurate.
3
Indeed, it would be difficult to do so given the ambiguity of the phrase ―more mature markets.‖
4
The real problem is that both comments gave investors and analysts the impression that Twitter‘s
5
user engagement was strong and increasing when Plaintiff claims the opposite was true. But the
6
Court already acknowledged that deception above.
7
In addition to these February 5 statements, Plaintiff also challenges Noto‘s statement
8
during Twitter‘s Q1 2015 Earnings Call on April 28 that ―DAU to MAU ratios in the quarter were
9
similar to what they were‖ at Analyst Day. Id. ¶ 102. At Analyst Day, Twitter reported a 48%
DAU to MAU ratio in its top 20 markets. Id. ¶ 85(j) (see chart). By contrast, Plaintiff claims that
11
United States District Court
Northern District of California
10
the DAU to MAU ratio in Q1 2015 was 44%, meaning that the ratio had declined since Analyst
12
Day and was not in fact ―similar‖ as Noto claimed. Id. ¶ 102(c). Defendants give several reasons
13
why this statement cannot support a 10(b) claim. First, Defendants claim that the term ―similar‖ is
14
not actionable because it is not capable of objective verification. ECF No. 91-1 at 23. In essence,
15
Defendants are arguing that Noto‘s statement is ―mere puffing.‖ But Defendants have not
16
identified any case that categorizes the word ―similar‖ as an ―unspecific assertion[] of corporate
17
optimism.‖ Cement & Concrete Workers Dist. Council Pension Fund v. Hewlett Packard Co., 964
18
F. Supp. 2d 1128, 1138–39 (N.D. Cal. 2013). Indeed, in Cement, on which Defendants rely, the
19
court defined puffery as ―vague statements of optimism like ‗good,‘ ‗well-regarded,‘ or other feel
20
good monikers.‖ Id. The Court does not agree with Defendants that ―similar‖ falls in this
21
category. In fact, the Ninth Circuit recently held a defendant‘s reassurances to investors that sales
22
were ―unchanged, or even growing, compared to previous quarters‖ was a ―concrete description,‖
23
not a ―‘feel good‘ optimistic statement[].‖ In re Quality Sys., Inc. Sec. Litig., 865 F.3d at 1144-
24
45. The same is true here.
25
26
Second, Defendants argue that even if ―similar‖ were a verifiable term, 48% is similar to
44%. But Plaintiff has alleged that a 4% difference is substantial and material, and on a motion to
27
28
29
1
dismiss, the Court must take their allegations as true.26 Compl. ¶ 102(a). Third, as discussed
2
above, Defendants identify what they claim are various problems with Plaintiff‘s calculations.
3
ECF No. 91-1 at 23-24 (explaining, for example, that ―the analyst report [from which Plaintiff
4
obtained the 44% figure] compares a ratio for the first three quarters of 2014 with a ratio for the
5
second quarter of 2015, and thus reflects a different comparison than that made by Noto in the
6
statement at issue‖). But, again, resolution of those issues should occur through discovery, not at
7
the motion to dismiss phase. Accordingly, the Court concludes that Plaintiff has stated an
8
affirmative misrepresentation claim based on Noto‘s April 28 statement that ―DAU to MAU ratios
9
in the quarter were similar to what they were‖ at Analyst Day. Id. ¶ 102.
Plaintiff also challenges Noto‘s comments during the April 28 Earnings Call about the
10
United States District Court
Northern District of California
11
success of Twitter‘s new product initiatives in driving user engagement. On that call, Noto said
12
that Twitter was ―seeing perhaps the most exciting results [from a new initiative called Recap] . . .
13
These tweets are not only seeing higher engagement, they‘re bringing people back to Twitter more
14
frequently. Importantly, the machine learning work we‘re doing for Recap is helping us to make
15
these algorithms better and driving continuous improvements in engagement.‖ Id. ¶ 103.27
16
Plaintiff argues these statements were misleading because Twitter later acknowledged that ―new
17
product initiatives ‗ha[d] not yet had a meaningful impact on growing our audience [i.e., MAU] or
18
participation [i.e., user engagement].‘‖ Id. ¶ 62. Again, Defendants argue that these statements
19
are ―generalized statements of optimism that constitute ‗non-actionable puffing.‘‖ Wozniak v.
20
Align Tech., Inc., 850 F. Supp. 2d 1029, 1036–37 (N.D. Cal. 2012) (quoting In re Syntex Corp.
21
Sec. Litig., 855 F. Supp. 1086, 1095 (N.D. Cal. 1994), aff'd, 95 F.3d 922 (9th Cir. 1996)). This
22
time, the Court agrees. Defendants comments about ―exciting results‖ and ―continuous
23
improvements‖ are the type of statements that Courts have deemed non-actionable, according to
24
26
25
26
Tellingly, Twitter seemed to think a 3% increase in the ratio was significant at Analyst Day,
when it projected that an increase of that magnitude in its top 20 markets would generate an
additional $500 million in revenue. Id. ¶¶ 39, 85(j).
27
27
28
Noto ―confirmed the success of the same new product initiative at an investor conference on
May 20, 2015, ―Dick talked about While YouWere Away on the earnings call and that we have
seen an improvement in engagement . . . .‖ Id. ¶ 103 n. 43.
30
1
the summary provided by a sister court:
2
Thus, for example, a court has held not actionable as ―mere puffery‖ statements
that ―[w]e are very pleased with the learning from our pilot launch,‖ ―so far we're
getting really great feedback,‖ and ―we are very pleased with our progress to date.‖
Wozniak v. Align Tech., Inc., 850 F. Supp. 2d 1029, 1036 (N.D. Cal. 2012).
Likewise, ―statements projecting ‗excellent results,‘ a ‗blowout winner‘ product,
‗significant sales gains,‘ and ‗10% to 30% growth rate over the next several years' ‖
have been held not actionable as mere puffery. In re Cornerstone Propane Partners,
L.P. Sec. Litig., 355 F. Supp. 2d 1069, 1087 (N.D. Cal. 2005); see also In re
Copper Mountain Sec. Litig., 311 F. Supp. 2d 857, 868–89 (N.D. Cal. 2004) (―runof-the-mill‖ statements such as ―business remained ‗strong‘ ‖ are not actionable
under § 10(b)); In re LeapFrog Enter., Inc. Sec. Litig., 527 F.Supp.2d 1033, 1050
(N.D. Cal. 2007) (vague statements such as ―This is going to be a very big second
half for us,‖ ―Our underlying sell-through at the retail level remained very strong
throughout the third quarter,‖ ―consumer demand for our learning products is more
vibrant than ever,‖ and ―We are pleased with our progress‖ were not actionable
under § 10(b)); City of Royal Oak Ret. Sys. v. Juniper Networks, Inc., 880 F. Supp.
2d 1045, 1064 (N.D. Cal. 2012) (statements that ―[b]oth Verizon and AT&T are
strong partners,‖ company has ―strong demand metrics and good momentum‖ and
―our demand indicators are strong, our product portfolio is robust‖ are unactionable
statements of corporate optimism).
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
In re Fusion-io, Inc. Sec. Litig., No. 13-CV-05368-LHK, 2015 WL 661869, at *14 (N.D. Cal. Feb.
12, 2015). Noto‘s statements about Twitter‘s new product initiatives are analogous. The Court
therefore grants the motion to dismiss as to these statements.28
Finally, Plaintiff challenges Defendants‘ positive statements about ad engagements as a
measure of user engagement, both to the SEC directly and to investors.29 In response to the SEC‘s
letter request that Twitter ―describe the alternative metric(s) you anticipate presenting in future
filings to explain trends in user engagement and advertising services revenue,‖ Compl. ¶ 113,
Twitter stated that it did include two user engagement metrics: ―changes in ad engagements and
changes in cost per ad engagement.‖ Id. ¶ 114. Twitter further explained that it ―internally tracks
changes in ad engagements . . . on the Twitter platform to monitor trends in user engagement . . .
and believes th[is] metric[] [is] helpful to investors to understand the same.‖ Id. Twitter also
24
28
25
26
27
28
The Court therefore does not address Defendants‘ alternative argument that these statements
were subject to the PSLRA‘s safe-harbor provision. ECF No, 91-1 at 34-35. Defendants do not
appear to argue that any of the other allegedly misleading statements are protected by the safe
harbor.
29
Plaintiff also discusses these statements in the context of their DAU omissions claim. For
simplicity of analysis (and because the Court has already concluded that Plaintiff state a claim
under their omission theory), the Court only discusses the ad engagement statements here.
31
1
touted its ad engagement numbers directly to investors. Id. ¶ 85(i)) (noting that in Q4 2014, for
2
example, ―[a]d engagements grew 70% year over year.‖ Plaintiff argues these statements were
3
misleading first because DAU, not ad engagements, was Twitter‘s primary user engagement
4
metric. They were also misleading, Plaintiff claims, because the ad engagements trend ―was
5
moving in the opposite direction from the concealed trend in actual user engagement (i.e., DAU).
6
ECF No. 85(i).
Defendants‘ response boils down to the following: even if DAU were the key user
8
engagement metric, ad engagements is another user engagement metric and it was not misleading
9
to accurately report ad engagement trends. ECF No. 91-1 at 39-40. The Court disagrees. It was
10
misleading for Defendants to rely on favorable ad engagement trends to describe or predict user
11
United States District Court
Northern District of California
7
engagement when DAU, Twitter‘s primary metric, was flat or declining.30 This is especially true
12
given the CWs‘ testimony that ―there is no direct correlation between advertising engagement and
13
MAU or DAU,‖ Compl. ¶ 77, because Ad Load could always be ―increased to compensate for the
14
lack of user (MAU) growth and user engagement (DAU) growth.‖ Id. ¶ 73. Therefore, even if ad
15
engagements could provide some insight into user engagement, Twitter‘s reliance on that metric
16
was misleading.31
17
C.
18
Finally, Plaintiff alleges that Defendants made false and misleading statements about
Claim Based on Affirmative Statements about Positive MAU Trends
19
Twitter‘s MAU trends. First, Plaintiff challenges Costolo‘s comments about MAU during the
20
February 5 Earnings Call, where he acknowledged lower than expected MAU growth at the end of
21
22
23
30
Defendants on some occasions described ad engagement as a ―monetization metric,‖ Compl. ¶
115, instead of a user engagement metric.
31
24
25
26
27
28
Both parties argue that Schueneman v. Arena Pharmaceuticals, Inc., 840 F.3d 698, 707 (9th Cir.
2016) supports their position on the ad engagement claim. In that case, the court held that
―Defendants may not have had a duty to disclose [an adverse study] had they not been
representing that [other] studies supported [the drug‘s] safety and therefore its likelihood of being
approved.‖ Id. This logic supports Plaintiff‘s omission theory here, where Defendants
represented that user engagement (measured by ad engagement) was improving when, according
to the DAU metric, it was not. Contrary to Twitter‘s argument, the Schueneman court never said
that the duty to disclose ―only‖ arose because Arena said ―all‖ animal studies supported the case
for FDA approval.
32
1
2014, but described signs of a rebound in early 2015:
2
We ended the quarter with 288 million monthly active users. We added 4 million
users this quarter and 47 million across 2014. There are quarter-specific factors that
impacted our net adds in Q4, which includes seasonality and a couple issues related
to the launch of iOS 8. We'll discuss that in more detail later in this call.
Importantly, I want to highlight that the user numbers we saw in January of this
year indicate that our MAU trend has already turned around and our Q1 trend is
likely to be back in the range of absolute net adds that we saw during the first three
quarters of 2014.
3
4
5
6
7
8
9
10
United States District Court
Northern District of California
11
12
Id. ¶ 88; ECF No. 92-7 at 3. When asked about the claims of ―acceleration‖ in MAU growth,
Costolo remarked, ―we‘re in a great place there‖ and attributed the Q1 turnaround to ―a
combination of seasonality or return to organic growth and the set of product initiatives [Twitter]
created to drive growth.‖ Compl. ¶ 88; ECF No. 92-7 at 8. Plaintiff argues these statements were
misleading because MAU had not ―turned around‖ or ―accelerat[ed]‖ by Q1 2015, but in fact was
―flat or declining leading up to and continuing through the Class Period.‖ Compl. ¶ 71.
Defendants reply that any statements about MAU could not be misleading because Twitter
13
14
15
16
17
18
19
20
21
22
23
24
25
consistently disclosed the Company‘s actual MAU numbers. ECF No. 91-1 at 36-37. During the
February 5 Earnings Call, Twitter acknowledged adding only 4 million users, Compl. ¶ 88, and
during the April 28 call, even Plaintiff describes Defendants as having ―disappointed investors by
lowering their MAU growth guidance for the coming quarter.‖ Compl. ¶ 105. Defendants also
point to Twitter‘s Q1 2015 Form 10-Q,32 which includes a chart showing a modest increase in
MAU in the first quarter of 2015, and eight quarters of previous MAU data. ECF No. 92-5 at 6
(showing 4 million new users in Q4 2014, and 14 million new users in Q1 2015). It was not
misleading for Defendants to say that MAU had turned around in Q1 2015 after a drop at the end
of 2014 when that is exactly what the data showed. Indeed, the CWs do not contest these
numbers. To the extent Plaintiff claims that Costolo‘s statement implied that the magnitude of the
turnaround or acceleration was greater than it actually was, no reasonable investor would have
relied on vague but generally positive remarks over the actual MAU data.
Plaintiff also claims, however, that Costolo‘s positive statements about MAU were
26
27
28
32
Twitter also references its 2014 Form 10-K, ECF No. 92-4 at 19, but that filing does not help
Defendants because it does not include MAU for Q1 2015, when MAU had allegedly ―turned
around.‖
33
misleading because any MAU growth included and was potentially driven by ―low quality users,‖
2
not ―organic‖ growth. Specifically, Twitter‘s MAU numbers were allegedly inflated by SMS or
3
registered third-party applications, by robot accounts, and by efforts of Twitter‘s marketing team.
4
Compl. ¶ 89(e), (f). This claim fails for two reasons. First, Twitter never claimed that its MAU
5
growth was wholly or even mostly organic. On the February 5 Earnings Call, for example,
6
Costolo attributed the Q1 2015 turnaround in MAU to ―a combination of seasonality or return to
7
organic growth and the set of product initiatives [Twitter] created to drive growth.‖ Compl. ¶ 88
8
(emphasis added). The fact that that Twitter did not also explain that some of its MAU growth
9
was not organic does not make Costolo‘s statement misleading. Second, Defendants disclosed the
10
existence of these ―low-quality‖ users. ECF No. 91-1 at 31. Indeed, Twitter always defined MAU
11
United States District Court
Northern District of California
1
to include ―SMS or registered third-party applications‖ and explained that the third-party ―activity
12
can cause our system to count the users associated with such applications as active users on the
13
day or days such contact occurs.‖ ECF No. 92-2 at 4 (Q2 2014 Form 10-Q).33 Similarly, Twitter
14
disclosed the fact some of its users were using ―robot accounts,‖ that is, fake or spam accounts that
15
―could have easily been created or bought.‖ Compl. ¶ 89(f). For example, in its Q2 2014 Form
16
10-Q, Twitter explained that it had ―performed an internal review of a sample of accounts and
17
estimate[d] that false or spam accounts represented less than 5% of our MAUs.‖ ECF No. 92-10
18
at 4. Finally, the fact that Twitter discussed its new product initiatives34 on various earnings calls,
19
33
20
21
22
23
24
25
26
27
28
Plaintiff argues that Twitter‘s ―Q1 2015 earnings presentation misleadingly included a
disclosure of the number of third-party automated MAUs from the prior quarter (i.e., as of
December 31, 2014) without an updated number of MAUs for Q1 2015, despite the
acknowledgement that the quarter was in fact impacted by automated users (―the calculations of
MAUs presented in presented in our earnings materials may be affected as a result of this
activity‖).‖ Compl. ¶ 107 n.47. The Court disagrees that this lack of update was misleading given
that, as Plaintiff acknowledges, Twitter did notify investors that Q1 had been impacted by these
automated users. Altayyar v. Etsy, Inc., No. 15CV2785AMDRER, 2017 WL 1157193, at *13
(E.D.N.Y. Mar. 16, 2017) (―The defendants cannot be held liable for failing to disclose something
that they disclosed.‖)
34
Mirroring their DAU-related claim, Plaintiff argues that Defendants statements about MAU
growth were misleading because ―new product initiatives were not having a meaningful impact on
driving MAU growth.‖ Compl. ¶ 89(g). The argument fails for the same reason: the challenged
statements are non-actionable puffery. For example, Costolo attributed an unidentified and
unverifiable part of Q1 2015 MAU growth to new product initiatives, stating that he ―hope[d] to
see the impact of these releases from both the growth and revenue perspective, over the coming
quarters.‖ ECF No. 92-7 at 5. Moreover, Noto stated during the same call that ―[a]lthough we
34
1
Compl. ¶ 88, means that the Company did not hide its attempts to improve DAU and MAU
2
numbers through marketing efforts. In other words, investors and analysts were on notice about
3
all three drivers of Twitter‘s allegedly low quality MAU growth. A lack of detailed information
4
on the numbers of these types of users during the Class Period does not itself make Defendants‘
5
statements about MAU growth misleading.
Scienter35
6
D.
7
Under the core operations doctrine, ―[a]llegations regarding management‘s role may help
8
satisfy the PSLRA scienter requirement in three circumstances:‖
9
First, the allegations may be viewed holistically, along with other allegations in the
complaint, to raise a strong inference of scienter under the Tellabs standard.
Second, the allegations ―may independently satisfy the PSLRA where they are
particular and suggest that defendants had actual access to the disputed
information,‖ as in Daou and Oracle. Third, in rare circumstances, such allegations
may be sufficient, without accompanying particularized allegations, where the
nature of the relevant fact is of such prominence that it would be ―absurd‖ to
suggest that management was without knowledge of the matter.
10
United States District Court
Northern District of California
11
12
13
Reese, 747 F.3d at 575–76.
14
15
Plaintiff has a strong argument that DAU and MAU were so integral to Twitter‘s success
that it would be ―absurd‖ to argue that Defendants Costolo and Noto36 were unaware of trends in
16
those two metrics during the class period. At Analyst Day, Twitter described its DAU to MAU
17
ratio as one of its five ―major growth drivers.‖ Id. ¶ 85(j). After Analyst Day, Costolo
18
emphasized the importance of DAU to investors and analysts, even more so than MAU:
19
20
21
22
don‘t expect the product launches and tests announced over the last two weeks [to] have a
meaningful impact on Q1 user growth, we‘re hopeful that these product initiatives will contribute
in subsequent quarters.‖ Id. at 6. In other words, even with respect to Q1 2015, Defendants
warned investors to be cautious about the effect of the Company‘s new product initiatives.
35
23
24
25
26
27
28
Defendants address scienter separately for each of the challenged statements. The Court does
not analyze scienter regarding Defendants‘ affirmative MAU statements, which it found were not
misleading, and considers scienter for the remaining statements together because they all require
knowledge of the same set of facts.
36
If the scienter requirement is met as to the Individual Defendants, it is met as to Twitter because
Defendants do not argue that Noto or Costolo were acting outside the scope of their apparent
authority. In re ChinaCast Educ. Corp. Sec. Litig., 809 F.3d 471, 476 (9th Cir. 2015) (―The
scienter of the senior controlling officers of a corporation may be attributed to the corporation
itself to establish liability as a primary violator of § 10(b) and Rule 10b–5 when those senior
officials were acting within the scope of their apparent authority.‖).
35
1
[Q:] Should people still focus on monthly active users? Is that what you look at on
the inside, is that the number-one metric you want to drive growth on?
2
[A:] Our number-one priority is to strengthen the core and make Twitter an
increasingly daily use case for the people who come to Twitter . . .
3
Id. ¶ 33 (February 5, 2015 Business Insider Interview with Costolo) (emphasis added). Similarly,
5
during a March 3, 2015 investor conference, Noto stated that Twitter‘s ―goal is to build the largest
6
daily audience that we can‖ in response to a question about the ―real incremental driver of user
7
growth.‖ ECF No. 106-9 at 4. The CWs confirmed the primacy of DAU during the Class Period
8
throughout the Company. E.g., id. ¶ 71 (―[T]he most important metrics to Twitter were MAU and
9
DAU, which were monitored closely by Twitter‘s management.‖). Even if the CWs statements
10
are unreliable as to the individual Defendants‘ state of mind, they reliably bolster the conclusion
11
United States District Court
Northern District of California
4
that DAU was Twitter‘s primary user engagement metric because of its direct impact on Twitter‘s
12
ability to generate revenue. Finally, after the end of the Class Period, Defendants stated
13
specifically that DAU was the ―the best encapsulation of engagement‖ and ―the best way to
14
quantify the impact of engagement.‖ id. ¶ 85(b). Viewing the evidence holistically, Plaintiff has
15
plausibly alleged that, from Analyst Day onward, Twitter consistently relied on DAU as its key
16
user engagement metric.
17
Given these allegations, it would be ―absurd‖ for Noto and Costolo not to have been aware
18
of adverse DAU trends during the class period,37 or to have misunderstood the interaction between
19
DAU and MAU.38 The DAU metric, the ―information [the Court] imput[es] to [Defendants,] is
20
37
21
22
23
Defendants say Plaintiff fails to ―articulate which facts Defendants must have known.‖ The
Court disagrees. Taking the facts in the light most favorable to Plaintiff, Defendants should have
known ―that engagement on the platform was decreasing throughout the Class Period,‖ Compl. ¶
128, and that DAU, Twitter‘s ―primary engagement metric‖ ―had declined significantly‖ during
the Class Period, id. ¶ 53.
38
24
25
26
27
28
Long before the Class Period began, Twitter acknowledged the connection between MAU and
DAU in its 2013 Registration Statement, explaining that ―[t]o the extent our user growth rate
slows, our success will become increasingly dependent on our ability to increase levels of user
engagement.‖ Id. ¶ 23. The limited utility of MAU without DAU was confirmed by multiple
CWs and reported by industry analysts. Id. ¶ 76 (CW-10) (―‗MAU was a terrible metric‘ and there
is no way you could ‗judge the health of the company using MAU alone.‘‖); Id. ¶ 121 (Tech
analyst Jan Dawson explained that ―[m]onthly usage metrics [MAUs] indicate very little about
true engagement on a platform, because using an app every 30 days isn‘t that much different from
never using it at all. For social and communication apps, the key is daily usage because it tells
you how people are really engaging.‖). Defendants cannot credibly claim they were unaware of
36
1
fundamental to operations of [Twitter‘s] business.‖ Reese, 747 F.3d at 576; compare Zucco
2
Partners, LLC, 552 F.3d 1001 (refusing to assume that defendants were aware that ―certain costs
3
had been erroneously capitalized because either a portion of such costs did not qualify for
4
capitalization . . . or the project itself did not qualify as internaluse software‖ as those facts were
5
not ―especially prominent‖).
6
The ―actual access‖ analysis also supports scienter. During Twitter‘s Q1 2015 Earnings
7
Call on April 28, 2015, in response to a question about how analysts ―should be thinking about
8
monitoring engagement,‖ Noto said:
9
In terms of engagement metrics, there‘s a lot of different metrics that we look at
internally. There‘s not one metric for engagement. And so I can give you a sense of
some of them and quite frankly, we would like to be able to give you more
visibility on this, but there‘s just a number of different measurements. DAU is one
measurement of engagement.
10
United States District Court
Northern District of California
11
12
Id. ¶¶ 42, 103. Declining to report specific DAU numbers, Noto explained: ―It‘s a measurement
13
that is dependent by market and you can have mixed shifts, so it could be a little bit misleading.‖
14
Id. Although Noto did not provide ―specific‖ data about DAU (in fact, he did the opposite),
15
Reese, 747 F.3d at 576, his statements still strongly imply that he had access to the disputed
16
information. How could Noto have known that DAU was ―dependent by market‖ or claim that it
17
could be ―misleading‖ without knowledge of the DAU numbers? Similarly, by claiming on the
18
April 28 call that ―DAU to MAU ratios in the quarter were similar to what they were‖ at Analyst
19
Day, id. ¶ 42, Noto arguably ―bridged the scienter gap [] by referencing the data directly.‖ Reese,
20
747 F.3d at 572 (citing S. Ferry, 542 F.3d at 783) (internal alterations omitted).39 As the Ninth
21
22
23
24
25
26
the interplay between these two metrics. With regard to ad engagements, Plaintiff has not
provided any evidence outside of the CWs‘ testimony that Noto or Costolo knew that ―linking user
engagement with ad engagement was misleading because ―Ad Load could always be ―increased to
compensate for the lack of user (MAU) growth and user engagement (DAU) growth.‖ Id. ¶ 73.
Nonetheless, even if Defendants were unaware that, as CW-11 stated, ―there is no direct
correlation between advertising engagement and MAU or DAU,‖ id. ¶ 110(c), so long as
Defendants knew that DAU and ad engagement trends were moving in opposite directions,
Defendants ―had reasonable grounds to believe‖ their favorable ad engagement statements were
misleading
39
27
28
These statements by Defendants distinguish this case from ones like Bodri v. GoPro, Inc., No.
16-CV-00232-JST, 2017 WL 1732022, at *13 (N.D. Cal. May 1, 2017), where the plaintiff alleged
nothing more than ―generalities about management‘s access to data.‖
37
1
Circuit has explained, an assertion that Defendants were unaware of an alleged issue can be
2
―directly contradicted by the fact that [they] specifically addressed it in [their] statement[s].‖
3
Reese, 747 F.3d at 572. By making ―detailed factual statement[s], contradicting important data to
4
which [the Individual Defendants] had access, a strong inference arises that [they] knowingly
5
misled the public as to its clear meaning.‖ Id. Here, Defendants emphasized the importance of
6
daily use and told investors that user engagement trends were positive. Because Defendants made
7
those factual statements without disclosing adverse DAU trends, and indeed, while refusing to
8
acknowledge the importance of DAU when asked for a measure of user engagement, ―a strong
9
inference arises that [they] knowingly misled the public‖ what was really happening with user
engagement at Twitter in 2015. Indeed, there is nothing to suggest that the bad news Defendants
11
United States District Court
Northern District of California
10
disclosed on July 28, which included data from the Class Period, had come as a surprise to Noto
12
and Costolo. On the contrary, according to the Vanity Fair article, the 2Q 2015 call represented a
13
conscious decision to ―come clean‖ about lagging user growth and engagement. Id. ¶ 78. 40
14
Plaintiff has also alleged a number of additional facts that support scienter. First, the
15
timing of Defendants‘ statements is suggestive of scienter. Defendants‘ evasive responses to
16
questions about DAU came just a few months after their aggressive user growth and engagement
17
projections at Analyst Day. Given that Twitter had just committed to the ―operational goal of
18
building the world‘s largest daily audience,‖ id. ¶ 31, Defendants ―had every reason to review‖
19
DAU and otherwise track user engagement. Reese, 747 F.3d at 571. When DAU trends did not
20
40
21
22
23
24
25
26
27
28
Defendants argue that the more plausible rationale for Twitter‘s decision not to disclose DAU
was its concern that DAU ―could be a little bit misleading‖ to investors and analysts because of
variances across its markets. Id. ¶¶ 88, 103. But Defendants do not explain why, just months after
the close the Class Period, DAU transformed from a misleading metric to one ―[t]he one
engagement metric that we look at holistically.‖ Id. ¶ 85(b). Nor do Defendants offer any reason
why they could not have continued to provide quarterly updates on the DAU to MAU ratio
throughout the Class Period, but broken down by market, if that was what rendered DAU
misleading. Indeed, as Plaintiff notes, many of Twitter‘s social media competitors reported DAU
on their platforms. Id. ¶ 121 (listing Facebook, LinkedIn, Snapchat, Instagram, and others).
Defendants dispute that these companies were ―peers‖ and argue that they define DAU differently
than Twitter. ECF No. 104 at 15. But those disputes are appropriately resolved on summary
judgment, not a motion to dismiss. Particularly given that the actual DAU numbers would have
upset Twitter‘s narrative of explosive growth, the Court finds Plaintiff‘s explanation for the
omission of granular DAU data (and corresponding focus on ad engagements) to be the most
plausible explanation for Defendants‘ statements.
38
1
match the Analyst Day projections in 2015, Defendants ―also had a clear motive for omitting
2
information about‖ adverse DAU trends. Id. Also related to timing is the fact that only a few
3
months elapsed between Twitter‘s first misleading statements on February 5 and the Company‘s
4
disclosures on July 28. In light of Twitter‘s consistent monitoring of user growth and user
5
engagement trends, this temporal proximity ―bolsters the inference of scienter.‖ Reese, 747 F.3d
6
at 575 (finding that a separation of ―three to six months between ―the statements and the
7
contradictory disclosure‖ can contribute to a finding of scienter ―given the relatively constant,
8
long-term nature of the information‖).
Plaintiff also claims that the departure of multiple senior Twitter officials supports an
10
inference of scienter. Costolo, Strickler, and other ―key executives‖ were ―forced‖ from their
11
United States District Court
Northern District of California
9
positions or resigned in the weeks before and after the July 28 disclosures. Id. ¶ 129-33. Multiple
12
CWs stated that, internally, Twitter‘s employees ―understood that Costolo [had been] pushed out
13
of the Company due to a lack of user growth.‖ Id. ¶ 67-67. Likewise, the Vanity Fair article
14
describes how Defendants felt they had to ―come clean‖ about stagnant user growth and
15
engagement, and that it was those disagreements that led, in part, to at least Strickler‘s departure.
16
Id. ¶ 89(c). Accordingly, because they are ―accompanied by additional evidence of [D]efendat[s‘]
17
wrongdoing,‖ these resignations or terminations constitute evidence of scienter. In re Downey
18
Sec. Litig., No. CV 08-3261-JFW(RZX), 2009 WL 2767670, at *13 (C.D. Cal. Aug. 21, 2009).
What does not support scienter by the individual Defendants is the testimony of the CWs.41
19
20
To contribute to a finding of scienter, CW testimony must pass two tests. First, the CWs must be
21
―described with sufficient particularity to establish their reliability and personal knowledge.‖
22
Zucco Partners, 552 F.3d at 995. Second, the CWs‘ statements must ―themselves [be] indicative
23
of scienter.‖ Id. The CWs do not pass the first hurdle; although they are described with enough
24
specificity to be reliable, the allegations in the Complaint do not permit the Court to conclude that
25
any CW had personal knowledge of the individual Defendants‘ state of mind.
26
Plaintiff does provide sufficient background on each CW. For example, Plaintiff provided the job
27
41
28
The exception is the CW testimony that supports the core operations theory, described above.
39
1
title of CW-1, the dates he was employed, a description of the projects he worked on that relate to
2
this case, and a summary of meetings he attended where user growth and engagement issues were
3
discussed. Compl. ¶ 67. The Complaint contains the same type of information for each of the
4
CWs, which is sufficient to establish their reliability. See In re Quality Sys., Inc. Sec. Litig., 865
5
F.3d 1130, 1144-45 (9th Cir. 2017). There are no allegations, however, demonstrating personal
6
knowledge. Critically, none of the CWs report communicating directly with Noto or Costolo. In
7
their opposition brief, Plaintiff claims that CW-2 described Tea Time meetings where Costolo was
8
present. ECF No. 94 at 25. But in the Complaint, CW-2 refers only to ―senior management,‖ not
9
to Costolo specifically. Compl. ¶ 68.42 Vague references to ―management,‖ or ―the Company‖ are
not sufficient to raise a strong inference of scienter for Noto or Costolo specifically, despite the
11
United States District Court
Northern District of California
10
corroborative nature of the CWs‘ testimony.
Zucco is instructive on this point. In that case, one CW stated that the defendant ―had to
12
13
have known what was going on with respect to the Company‘s inventory accounting
14
manipulation,‖ while another claimed that ―certain project managers knew‖ about the accounting
15
fraud. 552 F.3d at 998. The court held that ―[t]hese generalized claims about corporate
16
knowledge are not sufficient to create a strong inference of scienter, since they fail to establish that
17
the witness reporting them has reliable personal knowledge of the defendants‘ mental state.‖ Id.
18
The same is true here. Therefore, although the CWs general knowledge of the Company and its
19
metrics supports Plaintiff‘s core operations theory, their statements do not separately give rise to
20
any inference of scienter.
21
Defendants also emphasize the ―absence of allegations of relevant stock sales‖ during the
22
Class Period. ECF No. 91-1 at 22. However, as Plaintiff notes, their Complaint does not rely on
23
allegations of an improper financial motive to demonstrate scienter, nor does it reference stock
24
sales. Rather, Plaintiff‘s claim that Defendants were motivated by an attempt to live up to the
25
overly optimistic promises made at Analyst Day. The Court is therefore unwilling to ―draw a
26
27
28
42
Nor does the Vanity Fair article say that Costolo attended the Tea Time meetings. It merely
reports only that ―under Costolo, part of the Tea Time ritual included a show-and-tell to the
employees about the current state of the business,‖ including MAU. ECF No. 92-13 at 6.
40
negative inference from the absence of stock sales that benefitted the defendant chief executive
2
officer.‖ Sharenow v. Impac Mortg. Holdings, Inc., 385 F. App‘x 714, 717 n.2 (9th Cir. 2010)
3
(explaining that the defendants were allegedly motivated by ―wanting to keep the company going
4
to get large salaries than by making sales of their modest stock holdings‖); No. 84 Employer-
5
Teamster Joint Council Pension Tr. Fund v. Am. W. Holding Corp., 320 F.3d 920, 944 (9th Cir.
6
2003) (―Scienter can be established even if the officers who made the misleading statements did
7
not sell stock during the class period. . . . In other words, the lack of stock sales by a defendant is
8
not dispositive as to scienter.‖); compare In re Rigel Pharm., Inc. Sec. Litig., 697 F.3d 869, 884
9
(9th Cir. 2012) (finding that the defendants‘ lack of stock sales was weighed against a strong
10
inference of scienter because it was ―inconsistent with Plaintiff's theory that financial motive
11
United States District Court
Northern District of California
1
establishes scienter‖) (emphasis added). Here, Plaintiff‘s theory of the case is that Defendants felt
12
pressure to live up to the targets announced at Analyst Day, not that they sought personal financial
13
gain.
In sum, the core operations doctrine, together with Plaintiff‘s other allegations, create a
14
15
strong inference of scienter with respect to the misleading statements identified above.
16
E.
17
Because Plaintiff adequately pleaded a section 10(b) violation, the Court denies the motion
18
Control Person Liability
to dismiss as to Plaintiff‘s claim for control person liability under section 20(a).
CONCLUSION
19
The Court grants the motion to dismiss with respect to Plaintiff‘s claims based on
20
21
Defendants‘ affirmative statements about MAU, but denies it as to Plaintiffs DAU omissions
22
///
23
///
24
///
25
///
26
///
27
28
41
1
theory,43 and Plaintiff‘s claims based on Defendants‘ affirmative statements about positive user
2
engagement trends.44
IT IS SO ORDERED.
3
4
Dated: October 16, 2017
______________________________________
JON S. TIGAR
United States District Judge
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
43
26
The Court rejects the theory that the omission was misleading because DAU was Twitter‘s
primary user engagement metric.
27
44
28
The Court rejects any affirmative claims based on statements about Defendants‘ new product
initiatives, based on the statement about a 3% increase in timeline views, or based on the 50%+
statement (although the latter two still support the omissions claim).
42
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?