ORDER DENYING MOTION FOR CLASS CERTIFICATION

Re: Dkt. Nos. 261, 254, 257, 306

RONALD M. WHYTE, District Judge.

Plaintiffs move to certify a class of Facebook users pursuant to Federal Rules of Civil Procedure 23(a) and 23(b)(3). Dkt. No. 261. In addition to plaintiffs' motion for class certification, there are several related motions pending before the court: Facebook's motion to strike the expert opinion of Forrest Vickery, Facebook's motion to strike the expert opinion of Edward Pscheidt, and plaintiffs' motion to strike the expert testimony of Matt Jones. Dkt. Nos. 254, 257, 306. The court heard argument on January 29, 2016. Facebook's motion to strike the expert opinion of Edward Pscheidt is denied as moot because plaintiffs do not rely on Mr. Pscheidt's opinions. For the reasons explained herein, the court denies plaintiffs' motion for class certification and plaintiffs' motion to strike the testimony of Matt Jones. As the court does not reach the parties' arguments with respect to damages, Facebook's motion to strike the expert opinions of Forrest Vickery is denied as moot.

I. BACKGROUND

Facebook provides social networking services to its users free of charge. Dkt. No. 227, TAC ¶¶ 10, 11. Facebook users must register on the website using their real names. Id. ¶ 12. After registering, users may then post other personal information to a Facebook profile page. Id. ¶ 13. Facebook generates revenue through advertising. Id. ¶ 18. Facebook is able to target ads to particular demographics using the personal information provided by its registered users. Id. ¶¶ 26-27. Targeted advertising is valuable to Facebook's advertisers. Id. ¶¶ 26-27, 63, 65. According to Facebook, the "ad targeting is done entirely anonymously. If advertisers select demographic targeting for their ads, Facebook automatically matches those ads to the appropriate audience. Advertisers only receive anonymous data reports." Id. ¶ 28 (quoting Facebook's Privacy Guide). Plaintiffs allege that Facebook has "consistently and adamantly promised" not to share any user's specific identity or personally identifiable information ("PII") with its advertising partners. Id. ¶¶ 32, 27-31.

According to plaintiffs, Facebook's business model represents a bargain between Facebook and its users: "Facebook users provide their valuable PII to Facebook in exchange for access to facebook.com and for Facebook's promises not to disclose their PII to third parties without consent." Id. ¶¶ 16, 62, 67. Plaintiffs claim that Facebook violated its own policies and promises by disclosing users' "sensitive personally identifiable information" to advertisers. Id. ¶¶ 1-5. Specifically, plaintiffs allege that when users clicked on Facebook ads, Facebook sent "referer headers" to the advertisers that would allow the advertiser to identify the user who clicked the ad. Id. ¶¶ 34-39.

When an Internet user navigates to a website by clicking a link, the user's browser may send a "referer header" to the destination website's server. Dkt. No. 286, Jones Decl. ¶ 6. A "referer header" typically contains the URL of the website the user was visiting when he or she clicked the link, but Facebook may choose to include other information in the referer header as well. Id. Each Facebook user account is tied to a unique user ID number or username. TAC ¶ 14; Jones Decl. ¶¶ 4-5. Before July 2010, "if a Facebook user navigated the Facebook website in a particular way and clicked on a third-party advertisement linked to an external page, the resulting referer header may have included a user's Facebook username or user ID." Jones Decl. ¶ 9. A referer header containing a user ID or username from a referring Facebook profile page is not enough to identify the Facebook user who clicked on the ad; the same referer header would be generated for any user visiting that particular Facebook profile page, not just a user visiting his or her own Facebook profile page. Id.

During the same timeframe, however, Facebook may have "included a particular string of characters in users' profile page URLs (`ref=profile')" if a "user took certain steps to navigate to his or her own profile page." Jones Decl. ¶ 10 (emphasis added). As a result, some referer headers sent to external advertisers included "both the `ref=profile' string and the user ID or username of the user who clicked the ad." Id. ¶ 11; see also TAC ¶¶ 35-39. Facebook acknowledges that "in theory," a recipient of both the "ref=profile" string and a user ID or username would be able to identify the Facebook user who clicked on the ad. Jones Decl. ¶ 11. Plaintiffs assert two causes of action based on the transmission of these referer headers—breach of contract and fraud. TAC ¶¶ 58-73.

When a Facebook user clicks on an ad, the user is first sent to an "interstitial redirector," an intermediary webpage hosted by Facebook that records data relating to the user's ad click before directing the user's browser to the advertiser's targeted webpage. Dkt. No. 290, Hung Decl. ¶¶ 13-15. Plaintiffs rely on Facebook's ad click data to define the putative class. Ms. Marfeo seeks to represent a class of "[a]ll Facebook users in the United States who, at any time after May 28, 2006, clicked on a third-party advertisement displayed on Facebook.com configured to redirect the user to an external website, and which resulted in Facebook's capture of a referer header containing that user's Facebook userID and/or username." Dkt. No. 261 at 2.

Fact discovery for class certification closed on July 1, 2015, and expert discovery relating to class certification closed on October 7, 2015. See Dkt. No. 236. Facebook filed a motion to dismiss for lack of standing, which was granted in part, leaving Wendy Marfeo as the sole named plaintiff in this case. See Dkt. No 343.

II. MOTION TO STRIKE TESTIMONY OF MATT JONES

Plaintiffs move to strike paragraphs 15-19 of the Declaration of Matt Jones, which relate to operation of the Quickling tool, or in the alternative, move to reopen discovery on the topic of Quickling. Federal Rule of Civil Procedure 26(a)(2) requires the disclosure of expert witness. Federal Rule of Civil Procedure 26(e)(1) requires that a party supplement or correct its discovery response "in a timely manner if the party learns that in some material respect the disclosure or response is incomplete or incorrect, and if the additional or corrective information has not otherwise been made known to the other parties during the discovery process or in writing." Under Federal Rule of Civil Procedure 37(c)(1), a party who fails to provide information or identify a witness as required by Rule 26(a) or (e) is "not allowed to use that information or witness to supply evidence on a motion, at a hearing, or at a trial, unless the failure was substantially justified or is harmless." Plaintiffs argue that Facebook did not disclose Mr. Jones as an expert witness and did not update its discovery responses to sufficiently disclose Quickling. The court is not persuaded that Mr. Jones offers undisclosed expert testimony or that Facebook comply with its discovery obligations. Therefore, the court declines to strike the paragraphs 15-19 of the declaration, and plaintiffs' motion is denied.

A. Nature of Testimony

The court has reviewed paragraphs 15-19 and finds that they contain lay testimony based on factual knowledge acquired by Mr. Jones during his time as a Facebook software engineer. See Dynetix Design Sols., Inc. v. Synopsys, Inc., No. C 11-5973 PSG, 2013 WL 4537838, at *6 (N.D. Cal. Aug. 22, 2013) (percipient fact testimony on how source code works permitted from employee who worked on design and delivery). Mr. Jones claims to have personal knowledge of all facts in his declaration, except as otherwise noted. Paragraphs 15-19 contain no disclaimer as to personal knowledge. The paragraphs describe how the Quickling tool works.

Plaintiffs argue that these paragraphs contain expert opinion testimony because they include sentences describing how Facebook users "would have spent time" and how Quickling "would" or "could" have functioned. The court finds these sentences to be admissible lay opinions under Federal Rule of Evidence 701 because they are based on Mr. Jones' own perception of how the Quickling tool works and do not "provide specialized explanations or interpretations that an untrained layman could not make if perceiving the same acts or events." Laser Design Int'l, LLC v. BJ Crystal, Inc., No. C03-1179 JSW, 2007 WL 735763, at *3 (N.D. Cal. Mar. 7, 2007); Fresenius Med. Care Holdings, Inc. v. Baxter Int'l, Inc., No. 597, 2006 WL 1330002, at *3 (N.D. Cal. May 15, 2006) (permitting lay opinion testimony "premised on his personal knowledge regarding the machine and the way that it operates" but excluding opinion on how product compared to patent).

B. Facebook's Discovery Obligations

Aside from their contention that Facebook was required to disclose Mr. Jones as an expert, plaintiffs argue that Facebook had a duty to update its initial disclosures to identify Mr. Jones as a witness with knowledge of Quickling. Plaintiffs, however, acknowledge that Facebook's initial disclosures identify Mr. Jones as a witness on the "Facebook website and referrer headers." See Dkt. No. 306 at 9.

Plaintiffs also contend that Facebook was obligated to update its responses to Interrogatory Nos. 5, 6, 7, and 21 with details about Quickling. Interrogatory No. 5 is conditional, requiring a response only if Facebook contends that "no REFERRER HEADERS containing MARFEO's Facebook User ID OR Username OR other identifying information were transmitted to any PERSON." Dkt. No. 315-8 at 14. Facebook responded that it was "not aware of information sufficient to determine whether Plaintiff's ad clicks actually resulted in her browser's transmission to advertisers' external websites of a referer header including a URL with her user ID or username and the ref=profile character set." Facebook listed several reasons that it could not determine whether Ms. Marfeo's personal information had bene transmitted to any advertiser, including that "Facebook's internal systems may have altered the URL contained within the historical ad click data such that the referer header actually transmitted was different." Id. at 15.

Interrogatory No. 6 seeks details, including how, when, for what reason, "Facebook embeds or caused to be embedded Facebook User IDs OR Usernames to third parties." Dkt. No. 315-7 at 1. Facebook's response to Interrogatory No. 6 incorporates Facebook's response to Interrogatory No. 7. Id. at 4. Interrogatory No. 7 seeks identification of the total number of referer headers sent to Facebook advertisers that contained the username or user ID of the person who clicked on an advertisement. Id. Facebook responded that it did not know the total number. Id. at 6. Facebook listed several reasons that it did not know the total number, including that "Facebook's internal systems may have altered the URL contained within the historical ad click data such that the referer header actually transmitted was different." Id. Although Quickling is not identified by name, the court is not persuaded that Facebook violated a duty to supplement or correct its responses to Interrogatory Nos. 5, 6, and 7 with the name "Quickling."

Interrogatory No. 21 seeks "all facts" upon which Facebook had unequivocally denied a request for admission. See Dkt. No. 315-6 at 27. Plaintiffs specifically cite Facebook's denials of plaintiffs' Requests for Admission Nos. 25-26, 38-30, 34, and 52-53 as triggering Facebook's duty to disclose the operation of Quickling. Facebook, however, did not respond with respect to any of these Requests for Admission on the basis that plaintiffs had exceeded the number of permitted Interrogatories. See Dkt. No. 315-6 at 7. Plaintiffs do not dispute that they did not challenge this objection or move to compel a response from Facebook.

Plaintiffs also argue that Quickling was not sufficiently identified or explained in Facebook's document production or the deposition testimony of Facebook's witnesses, but plaintiffs do not identify specific document requests or deposition questions that went unanswered.

III. MOTION FOR CLASS CERTIFICATION

"Before certifying a class, the trial court must conduct a `rigorous analysis' to determine whether the party seeking certification has met the prerequisites of Rule 23." Mazza v. Am. Honda Motor Co., Inc., 666 F.3d 581, 588 (9th Cir. 2012) (quoting Zinser v. Accufix Research Inst., Inc., 253 F.3d 1180, 1186, amended 273 F.3d 1266 (9th Cir. 2001)). Under Rule 23(a), a plaintiff must demonstrate numerosity, commonality, typicality and adequacy of representation in order to maintain a class. Fed. R. Civ. P. 23(a)(1)-(4). If all four prerequisites of Rule 23(a) are satisfied, the court must then determine whether the class may be certified pursuant to one of Rule 23(b)'s three subsections. In this case, plaintiffs seek certification pursuant to Rule 23(b)(3), which allows class certification if the "court finds that the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy." Fed. R. Civ. P. 23(b)(3). "A party seeking class certification must affirmatively demonstrate his compliance with the Rule—that is, he must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact, etc." Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350-52 (2011); see also United Steel, Paper & Forestry, Rubber, Mfg. Energy, Allied Indus. & Serv. Workers Int'l Union, AFL-CIO, CLC v. ConocoPhillips Co., 593 F.3d 802, 807 (9th Cir. 2010) ("party seeking class certification bears the burden of demonstrating that the requirements of Rules 23(a) and (b) are met").

A. Rule 23(a) Prerequisites

Plaintiffs maintain that the Rule 23(a) requirements of numerosity, commonality, typicality and adequacy of representation are met; Facebook challenges plaintiffs' Rule 23(a) showing with respect to typicality and adequacy of representation.

1. Numerosity

A class must be so numerous that joinder of all members is impracticable. Fed. R. Civ. P. 23(a)(1). Plaintiffs seek to certify a nationwide class with millions of members, and Facebook does not challenge the numerosity of the proposed class. Plaintiffs have satisfied the numerosity requirement of Rule 23(a). See, e.g., Hanlon v. Chrysler Corp., 150 F.3d 1011, 1019 (9th Cir. 1998) (finding numerosity requirement "clearly satisfied" for nationwide class with millions of class members).

2. Commonality

Rule 23(a)(2) requires "questions of law or fact common to the class." Fed. R. Civ. P. 23(a)(2). Plaintiffs argue that there are several questions subject to common proof in this case, including Facebook's promise not to disclose users' personal information to advertisers and Facebook's alleged breach of that promise. Facebook contends that common questions do not predominate as required by Rule 23(b)(3), but Facebook does not challenge plaintiffs' showing of commonality under Rule 23(a). Given that "even a single common question will do" for Rule 23(a)(2) purposes, the court finds that plaintiffs have satisfied the commonality requirement. Dukes, 564 U.S. at 359 (citations omitted).

3. Typicality

A class representative's claims and defenses must be typical of the claims or defenses of the class. Fed. R. Civ. P. 23(a)(3). "The test of typicality `is whether other members have the same or similar injury, whether the action is based on conduct which is not unique to the named plaintiffs, and whether other class members have been injured by the same course of conduct.'" Ellis v. Costco Wholesale Corp., 657 F.3d 970, 984 (9th Cir. 2011) (quoting Hanon v. Dataproducts Corp., 976 F.2d 497, 508 (9th Cir. 1992)). Ms. Marfeo seeks to represent the class of "[a]ll Facebook users in the United States who, at any time after May 28, 2006, clicked on a third-party advertisement displayed on Facebook.com configured to redirect the user to an external website, and which resulted in Facebook's capture of a referer header containing that user's Facebook userID and/or username." See Dkt. No. 261 at 2. The parties do not dispute that Ms. Marfeo clicked on a Facebook advertisement that was configured to redirect Ms. Marfeo to an external website and which resulted in Facebook's capture of a referer header containing her userID and/or username.

Instead, Facebook contends that Ms. Marfeo's claims and defenses are not typical because her credibility is compromised by her felony embezzlement conviction and her related deposition testimony. Facebook does not explain how an unrelated felony conviction—or deposition testimony about such a conviction—renders Ms. Marfeo's claims atypical or creates any defenses unique to Ms. Marfeo's claims. The cases cited by Facebook involve credibility challenges directly related to the named plaintiff's claims. See Guido v. L'Oreal, USA, Inc., No. CV 11-1067 CAS JCX, 2012 WL 2458118, at *4 (C.D. Cal. June 25, 2012) (finding that discrepancies between plaintiff's deposition testimony and other evidence relating to claims "threaten to undermine her credibility, providing an alternative basis for the conclusion that her claims are atypical"); Fosmire v. Progressive Max Ins. Co., 277 F.R.D. 625, 633 (W.D. Wash. 2011) (finding that named plaintiff's "material misrepresentation upon her insurance application" provided defendant with an atypical defense against plaintiff's insurance claims); Drake v. Morgan Stanley & Co., No. CV 09-6467 ODW RCX, 2010 WL 2175819, at *5 (C.D. Cal. Apr. 30, 2010) (finding that named plaintiffs' wage claims against employer were subject to unique defenses based on plaintiff's unique employment background and lack of credibility regarding his own professional performance). The court is satisfied that Ms. Marfeo's claims in this case are typical of the proposed class.

4. Adequacy of Representation

Rule 23(a)(4) requires plaintiffs to show that the named plaintiffs will "fairly and adequately protect the interests of the class." Fed. R. Civ. P. 23(a)(4). Plaintiffs submit that they have retained experienced class counsel who successfully appealed the dismissal of plaintiffs' claims, and that there are no conflicts between absent class members and either the putative class representatives or counsel. Facebook does not challenge the adequacy of class counsel, but does challenge Ms. Marfeo's adequacy as class representative. First, Facebook argues that Ms. Marfeo's credibility is called into question by her felony embezzlement conviction and related deposition testimony. Second, Facebook argues that Ms. Marfeo is not sufficiently aware of the events in this case to represent the class. Third, Facebook argues that Ms. Marfeo testified in her deposition that her health would prevent her from traveling to California for trial. The court is not persuaded by any of these arguments.

a. Credibility

Honesty and credibility are relevant to the adequacy inquiry "because an untrustworthy plaintiff could reduce the likelihood of prevailing on the class claims." Harris v. Vector Mktg. Corp., 753 F.Supp.2d 996, 1015 (N.D. Cal. 2010) (citations omitted). To render a putative class representative inadequate, a credibility attack must be "so sharp as to jeopardize the interests of absent class members." Id. "This standard is difficult to satisfy." Mendez v. R+L Carriers, Inc., No. C 11-2478 CW, 2012 WL 5868973, at *14 (N.D. Cal. Nov. 19, 2012) (citing Cruz v. Dollar Tree Stores, Inc., No. 07-2050 S.C. 2009 WL 1458032, at *7 (N.D. Cal. May 26, 2009), modified in part, 270 F.R.D. 499 (N.D. Cal. 2010) ("inconsistency between his deposition testimony and statements in a declaration is not sufficient to impugn [plaintiff's] credibility")). "There is inadequacy only where the representative's credibility is questioned on issues directly relevant to the litigation or there are confirmed examples of dishonesty, such as a criminal conviction for fraud." Id. A criminal record, however "does not automatically disqualify a putative class representative." Kirkpatrick v. Ironwood Comm'ns, No. C05-1428JLR, 2006 WL 2381797, at *6 (W.D. Wash. Aug. 16, 2006).

Facebook challenges Ms. Marfeo's credibility based on her felony embezzlement conviction and her related deposition testimony. Plaintiffs do not dispute that Ms. Marfeo's nolo contendere plea resulted in a 2014 conviction for felony embezzlement in Rhode Island. The cases cited by Facebook, however, do not establish that a single felony embezzlement conviction is enough to disqualify Ms. Marfeo as class representative.

In Jamison v. First Credit Servs., Inc., for example, the district court found that plaintiff's "recent fraud conviction combined with the question of whether [plaintiff] pays the bill for the cellphone in question could lead the jury to focus on [his] credibility as opposed to the claims of absent class members." 290 F.R.D. 92, 105 (N.D. Ill. 2013). There is no similar issue of credibility that relates to Ms. Marfeo's class claims; as plaintiffs point out, the vast majority of the evidence in this case comes from Facebook's own data, rather than the testimony of Ms. Marfeo. The court notes that Facebook appears to agree that plaintiffs "rely exclusively on Facebook's internal `ad click logs' to establish which putative class members' clicks resulted in the alleged transmission of the challenged referer headers to advertisers." See Dkt. No. 285 at 2.

In Kirkpatrick, the district court found "no indication" that a single, ten-year-old, conviction for theft would have significant impact on a named plaintiff's ability to represent class, although the court disqualified another named plaintiff based on three recent felony convictions, at least two of which were for crimes "involving dishonesty." 2006 WL 2381797, at *6. Ms. Marfeo's sole felony conviction is relatively recent, but the court is not convinced that it would have a significant impact on her ability to represent the putative class in this case.

Nor is the court persuaded that Ms. Marfeo's allegedly false testimony about the conviction and the underlying incident disqualifies her as a class representative. See Mendez, 2012 WL 5868973, at *14 (rejecting efforts to disqualify representative based on alleged inconsistencies between representative's deposition testimony and other evidence in the case); cf. Kline v. Wolf, 702 F.2d 400, 403 (2d Cir. 1983) (affirming inadequacy based on class representative's admittedly false testimony and "refusal to answer proper discovery questions").

b. Knowledge of Litigation

Facebook faults Ms. Marfeo for her lack of knowledge about the case consolidation, the transfer of this case to the Northern District of California, and the appeal of the dismissal of plaintiffs' claims in this case. "The threshold knowledge required of the class representatives is low." DuFour v. Be LLC, 291 F.R.D. 413, 419 (N.D. Cal. 2013); see also Surowitz v. Hilton Hotels Corp., 383 U.S. 363, 372 (1966) (rejecting proposition that plaintiff "who is uneducated generally and illiterate in economic matters, could never under any circumstances be a plaintiff in a derivative suit brought in the federal courts to protect her stock interests"). A "party must be familiar with the basic elements of her claim, and will be deemed inadequate only if she is startlingly unfamiliar with the case. It is not necessary that a representative be intimately familiar with every factual and legal issue in the case." DuFour, 291 F.R.D. at 419 (citing Moeller v. Taco Bell Corp., 220 F.R.D. 604, 611 (N.D. Cal. 2004)); see also Parrish v. Nat'l Football League Players Ass'n, No. C 07-00943 WHA, 2008 WL 1925208, at *6-7 (N.D. Cal. Apr. 29, 2008) (finding that class representative's "conflicting and peculiar answers in response to questioning" relating to suit were "not alarming" in light of "his age and unfamiliarity with the legal process"); Alba v. Papa John's USA, Inc., No. CV 05-7487 GAF (CTX), 2007 WL 953849, at *9 (C.D. Cal. Feb. 7, 2007) (finding class representatives adequate "even if [they] do not understand the legal issues involved in their case" where defendants failed to demonstrate that representatives lacked knowledge regarding the facts in their case).

Having reviewed the deposition testimony submitted by the parties, the court is persuaded that Ms. Marfeo is familiar with the basic elements of her claim. See Dkt. No. 260-31 at 226:14-11 227:24 (testifying that Facebook's privacy policy was important to her and that Facebook "gave out my name. I wasn't paid for it and if I—it's valuable."); cf. Azoiani v. Love's Travel Stops & Country Stores, Inc., No. EDCV0790ODWOPX, 2007 WL 4811627, at *2 (C.D. Cal. Dec. 18, 2007), abrogated on other grounds by Bateman v. Am. Multi-Cinema, Inc., 623 F.3d 708, 714 (9th Cir. 2010) (finding representative inadequate where testimony established that not only had class representative "ceded all control to his counsel, but also that he had very little knowledge of the case to begin with"). Furthermore, the court is not startled by Ms. Marfeo's lack of familiarity with the litigation history of this action, particularly because Ms. Marfeo was reinstated as named plaintiff only one year ago—well after the consolidation of the cases, the transfer to this district, and the appeal of dismissal.

c. Ability to Travel

Ms. Marfeo's testimony on her ability to travel for trial is somewhat ambiguous. Compare Dkt. No. 287-12 at 189 ("I really can't be on a place for that amount of time or in a sitting position for that amount of time . . . I don't know how I would do it. . . . I'd be willing, but I don't know if I'm able if my condition's exactly how it is now."), with Dkt. No. 287-12 at 233 ("if my condition—my back and my leg is, you know—is able to travel and I could take someone and I could actually sit on the plane for that amount of time, absolutely."). However, Ms. Marfeo was able to sit for a full day deposition in this case, and the court is not persuaded that the health issues she describes are severe enough to disqualify her as a class representative. Cf. Krim v. pcOrder.com, Inc., 210 F.R.D. 581, 589 (W.D. Tex. 2002) (disqualifying representative where health had "already prevented his participating in at least one key hearing and representative testified that he did "not anticipate his health situation [would] significantly improve").

B. Rule 23(b)(3) Requirements

In addition to satisfying the Rule 23(a) requisites, plaintiffs must also demonstrate that "questions of law or fact common to class members predominate over any questions affecting only individual members" and that "a class action is superior to other available methods for fairly and efficiently adjudicating the controversy." Fed. R. Civ. Pro. 23(b)(3). Facebook argues that plaintiffs have not established the predominance of common questions, but does not challenge plaintiffs' showing that a class action is superior to other methods of resolving plaintiffs' claims.

1. Predominance

The predominance inquiry "tests whether proposed classes are sufficiently cohesive to warrant adjudication by representation." Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 623 (1997). This inquiry requires weighing the common questions in the case against the individualized questions, which differs from the Rule 23(a)(2) inquiry as to whether the plaintiff can show the existence of a common question of law or fact. See Dukes, 564 U.S. at 359. There is "`clear justification for handling the dispute on a representative rather than an individual basis' if `common questions present a significant aspect of the case and they can be resolved for all members of the class in a single adjudication.'" Mazza, 666 F.3d at 589 (quoting Hanlon, 150 F.3d at 1022. "Considering whether `questions of law or fact common to class members predominate' begins, of course, with the elements of the underlying cause of action." Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011).

Here, plaintiffs allege two causes of action under California law—breach of contract and fraud. Facebook argues that individual issues predominate with respect to the elements of 1) Facebook's breach or misrepresentation, 2) the injury to class members caused by any such breach or misrepresentation, and 3) class members' reliance on any such misrepresentation by Facebook. Because the court finds that the elements of breach and misrepresentation—which are central to plaintiffs' causes of action—will require individualized inquiry, the court does not reach Facebook's arguments with respect to the elements of damages and reliance.

Plaintiffs' claims for breach of contract and fraud rely on the same allegations: Facebook broke a promise by transmitting referer headers to advertisers that would allow advertisers to identify the user who clicked on the ad. The parties do not dispute that the Facebook's ad click data shows that ad clicks by putative class members resulted in referer headers containing either a userID or username and a "ref=profile" string. Facebook argues that the ad click data cannot serve as common proof of breach or misrepresentation because the data does not reliably reflect the content of the referer headers that were transmitted to third party advertisers. Facebook also argues that the ad click data cannot be common proof of breach or misrepresentation because it does not establish third party receipt or awareness of the referer headers.

The court has already addressed Facebook's second argument in its order on Facebook's motion to dismiss, determining that third party receipt or use is not required to establish breach of contract in this case. See Dkt. No. 343 at 8. The court is not persuaded by Facebook's citation to a dictionary of definition of "share" in further support of this argument. See Dkt. No. 285 at 16 n.11 ("share" is defined as "to partake of, use, experience, or enjoy with others" or "to grant . . . a share in," with a "share" meaning "the part allotted or belonging to one of a number owning together any property or interest"). Even under Facebook's proposed definition of "share," transmission of the information alone could constitute a "grant" or a "share" in plaintiffs' personal information— whether or not a third party advertiser accepted or made use of the grant from Facebook.

However, plaintiffs have not met their burden to show that the question of Facebook's transmission of the referer headers can be resolved by common proof. Plaintiffs contend that Facebook's ad click data is sufficient to show that users' information was transmitted to third party advertisers "in the ordinary course of how the Internet works." Dkt No. 324 at 8. Facebook, however, points to several reasons that the ad click data is not reliable evidence of transmission.

First, Facebook presents evidence that Facebook used a tool called "Quickling" to enable users' browsers load Facebook webpages more quickly. See Jones Decl. ¶ 15. When Quickling is enabled and a user navigates to another page, Quickling determines which portions of the first page do not need to be reloaded for display of the second page. Id. By loading only new portions of the second page, Facebook cuts down on browser load time. Id. During a Quickling session, if the new page does not need to be fully loaded, then URL does not change to the URL of the new page. Id. at 16. Facebook represents that Quickling was in widespread use between August 2009 and July 2010 and that Facebook users would have spent a significant amount of time in Quickling sessions. Id. at 15. Therefore, Facebook represents, REDACTED REDACTED Id. at ¶¶ 17-18.

Second, Facebook's digital forensics expert, James E. Hung, found evidence that many different privacy and security tools in use during the relevant time period were capable of modifying or removing referer information before it was sent to third party advertisers. Hung, Decl. ¶¶ 16-19. Mr. Hung tested historical versions of two programs, Privoxy and RefControl, and consulted historical documentation for a third program, Norton Internet Security. Id.

Third, Facebook's submits evidence that the Internet Explorer web browser would not have sent referer headers under certain circumstances. Facebook used two methods for redirecting a user's browser to the advertiser's target webpage during the relevant time period. Jones Decl. ¶ 13. Facebook's default method was the HTTP "302" redirect, in which Facebook's web server responded to a user's ad click by sending a standard HTTP response message using the "302" code to the user's browser. Id. The "302" code indicates that the browser should seek the requested webpage at a specified location. Id. In some circumstances, Facebook performed redirection by sending the user's browser a piece of software code in JavaScript, a programming language widely used on the Internet, to indicate that the user's browser should request a new webpage at a specified location. Id. Mr. Hung's testing shows that the versions of the Internet Explorer browser in use at the time "did not send referer headers after a JavaScript redirection was employed." Hung Decl. ¶ 26. "Thus, if a user were to click on an ad on Facebook using Internet Explorer and Facebook's interstitial redirector were to employ a JavaScript redirect, the user's browser would not have sent any referer header to the third-party advertiser's server, while Facebook would have received a normal referer value." Id. According to Facebook's former software engineer, Mr. Jones, REDACTED Jones Decl. ¶ 14.

Fourth, Facebook's expert opines that "companies and other organizations often employ web filtering proxies to protect their networks from sending information about their internal infrastructure and data to third parties," and that if a user accessed Facebook from a network running "a web proxy that was configured to block or alter referer information sent to third-party hosts," a third-party advertiser's server "would have been sent altered or no referer information, whereas Facebook would have been sent unaltered referer information." Hung Decl. ¶¶ 23-24.

Plaintiffs object that Facebook offers no proof that the transmission of a referer header was ever actually blocked or altered by any of these mechanisms. Specifically, plaintiffs point out that Mr. Jones does not indicate how widespread Facebook's use of Quickling was during the relevant time period—such as by describing how or when a Quickling session would begin or end. Similarly, plaintiffs object that Mr. Hung offers no evidence that privacy and security programs were in widespread use; he offered an estimated number of downloads for only one privacy tool— Privoxy—and made no showing that Privoxy had ever been downloaded by a putative class member. Plaintiffs further contend that Facebook has no evidence that a JavaScript redirect in Internet Explorer or a company's use of a web proxy server ever resulted in inaccurate ad click data.

It is plaintiffs' burden to establish the predominance of common questions. Plaintiffs' claims for breach of contract and fraud require proof that the referer headers containing putative class members' user ID or username and the `ref=profile' string were transmitted to third parties, and in order to certify the class, plaintiffs must show that its claims are predominantly subject to common proof. In order to avoid class certification, Facebook is not required to prove that transmission did not occur—Facebook must show only that the question is not subject to common proof. See Moore v. Apple Inc., No. 14-CV-02269-LHK, 2015 WL 7351464, at *8 (N.D. Cal. Nov. 20, 2015) ("Resolving whether there was an actual breach of a class member's contract requires individualized inquiries into whether the class member had a right to receive the missed text messages, therefore preventing the Court from certifying the class under Rule 23(b)(3)."). Facebook has produced unrefuted evidence that the questions of breach and misrepresentation cannot be resolved through Facebook's ad click data alone. Therefore, plaintiffs have not established that breach or misrepresentation can be established by common proof on a class-wide basis.

Plaintiffs argue that Facebook's own data may yet provide common proof on the questions of breach and misrepresentation with respect to any JavaScript redirects and the use of Quickling. Plaintiffs argue that Facebook previously informed plaintiffs that Facebook could not determine whether ad clicks redirected to an advertiser's Facebook page or a third party website, but later revealed that Facebook could make that determination. See, e.g., Dkt. No. 260-13 at 3 (Facebook's March 17, 2015 response to Interrogatory No. 7 stating that "Facebook cannot determine at this time whether a particular ad click directed internally or externally."). Plaintiffs thereby imply that some of the individualized questions that Facebook now raises may be capable of resolution through Facebook's internal data. Plaintiffs also cite to notes taken by Facebook's digital forensics expert during an April 16, 2015 telephone call with Facebook's team, in which Mr. Hung appears to quote Mr. Jones as saying that except for going to the advertiser, consulting Facebook's "code" would be the only way to tell what information was included in the referer header that was actually transmitted to the advertiser. Dkt. No. 260-29 at 3. Class certification is now closed, and plaintiffs' speculation on these issues does not establish that transmission is a common question of fact.

Plaintiffs also argue that putative class members could self-identify with respect to the use of any web proxies or privacy software. However, Facebook has not established that the question of transmission can be resolved by self-identification. As Facebook points out that Ms. Marfeo was unable to answer questions about her use of privacy or security software in her deposition. See Dkt. No. 287-12 at 107:3-7(Q. So you don't know — just to be clear, you don't know how your browser was configured to handle the transmission of referrer headers; is that correct? A. Yeah. I don't know — I don't — I have no idea.). Furthermore, plaintiffs do not explain how putative class members who accessed Facebook from a company or other network would have knowledge regarding the network administrator's decision on whether to use a web proxy server. Plaintiffs do not even suggest that class members could self-identify with respect to Facebook's internal mechanisms. Therefore, the court concludes that plaintiffs have not established that questions of breach or misrepresentation are capable of resolution on a class-wide basis.

IV. CONCLUSION

Because the court finds that individualized questions will predominate with respect to Facebook's alleged breach and misrepresentation, the court denies plaintiffs' motion for class certification.

The court has also filed an unredacted copy of this order that is accessible to case participants only. If either party believes that the redacted portions of this order disclose confidential information, the party must file an administrative motion to file under seal within 14 days of this order. The motion must be accompanied by an unredacted version of the order, filed under seal, highlighting the portions of the order that the party seeks to seal. The motion must also be accompanied by a declaration establishing that such portions of the order are sealable. No proposed order or redacted version of the order need be filed. If neither party moves to seal any part of this order within 14 days, the order will be made public in its entirety.

IT IS SO ORDERED.