Corley v. Google, Inc.

ORDER GRANTING MOTIONS TO SEVER Re: Dkt. No. 96 (No. 16-CV-00473-LHK) Re: Dkt. No. 20 (No. 16-CV-02553-LHK) LUCY H. KOH, United States District Judge Plaintiffs in these cases, Corley v. Google, Inc. (“Corley”), and Amaral v. Google, Inc. (“Amaral”),1 are 879 individuals who had Google Apps for Education accounts from November 1, 2010 to April 30, 2014. Plaintiffs allege that Google, Inc. (“Google”) violated the Wiretap Act by intercepting and scanning Plaintiffs’ emails. Before the Court are Google’s motions to sever, filed in the Corley and Amaral cases. ECF No. 19 (“FAC”); ECF No. 96 (“Mot.”); 16-CV-2553, ECF No. 20. With the Court’s permission, Public Citizen, Inc. filed an amicus brief opposing Google’s motion to sever. Having considered the submissions of the parties, the relevant law, and the record in this case, the Court GRANTS Google’s motions to sever. I. BACKGROUND A. Factual Background Google offers customers a “suite of integrated Google products or services” known as “Google Apps.” FAC ¶ 10. Google Apps for Education (“GAFE”) is a customized “version of Google Apps that Google offers to universities and other educational institutions ... for use by ... students, faculty, and staff.” Id. Among the products and services included in GAFE is Gmail, an email service. Id. GAFE users “receive email accounts and addresses with names that reflect the users’ Educational Institution affiliation, not an association with Google — e.g., jane.doe@ berkeley.edu.” Id. Plaintiffs aver that, from November 1, 2010 until April 30, 2014, Google “seann[ed] and processed] the content of every email received by or sent from a Google Apps for Education user account to develop sophisticated individual profiles for Commercial Purposes.” Id. ¶ 12. These actions allegedly violated the Wiretap Act, which generally prohibits the interception of wire, oral, or electronic communications. 18 U.S.C. § 2611(1). This is not the first time that GAFE users have brought a Wiretap Act claim against Google. In In re: Google Inc. Gmail Litigation (“Gmail”), 13-MD-2430 (N.D.Cal.), a multidistrict litigation assigned to the undersigned judge, nine plaintiffs asserted a Wiretap Act claim on behalf of GAFE and non-GAFE Gmail users. On October 26, 2013, plaintiffs in Gmail moved for “certification of ... four classes and three subclasses.” In re Google Inc. Gmail Litig., 2014 WL 1102660, *10 (N.D.Cal. Mar. 18, 2014). One such sub-class, the “Education Class,” sought to represent “[a]ll Google Apps for Education users who have, through their Google Apps for Education accounts, sent an email to an ‘@gmail. com’ address or have received an email.” Id. In denying class certification to the Education Class, the Court first noted that the Wiretap Act is not violated if either party to a communication consents to its interception. With this exemption in mind, the Court went on to note that, in order to create a Gmail account, all Gmail users — both GAFE and non-GAFE — had to agree to Google’s Terms of Service (“TOS”) and Privacy Policy. Id. at *3. From April 16, 2007 to March 1, 2012, Google’s TOS stated that it would “pre-screen, review, filter, modify, refuse or remove any or all Content from any Service”— including Gmail — in order to monitor “objectionable” content. Id. From 2010 to March 1, 2012, Google’s Privacy Policy stated that, “when you send email or other communications to Google, we may retain those communications in order to process your inquiries, respond to your requests and improve our services.” Id. at *4. After March 1, 2012, Google amended its TOS such that all Gmail users would now have to agree to “give Google ... a world [280] wide license to use ... and distribute [user] content.” Id. at *3. The Privacy Policy also changed on March 1, 2012, and “stated that Google could collect information that users provided to Google, cookies, log information, user communications to Google, information that users provide to affiliated sites, and the links that a user follows.” Id. In April 2014, after this Court’s Gmail order, Google amended its Privacy Policy to state that the email of GAFE and non-GAFE users would be analyzed. Matera v. Google, Inc., 15-CV-4062 (N.D.Cal), ECF No. 20-2 at 3 (“Our automated systems analyze your content (including emails) to provide you personally relevant product features, such as customized search results, tailored advertising, and spam and malware detection. This analysis occurs as the content is sent, received, and when it is stored.”). In addition, from 2011 to 2012, Google also maintained a “Privacy Center” website which stated that “Google scans the text of Gmail messages in order to filter spam and detect viruses. The Gmail filtering system also scans keywords in users’ email which are then used to match and serve ads.” 2014 WL 1102660, *6. From 2009 to 2014, Google maintained other webpages, blogs, and online tools to provide information on its interception and scanning practices. See id. at *4-*5. There was also media coverage of Google’s scanning practices, including New York Times, National Public Radio, and Washington Post articles and reports. Id. at *7. Aside from these generally applicable disclosures that were made to all Gmail users (both GAFE and non-GAFE), each educational institution appeared to have a unique privacy policy. As noted in Gmail, “the educational institutions with whom Google contracted were ... required to obtain the necessary authorizations from end users for Google to provide its services.” Id. at *6. These authorizations were to disclose, at the very least, the TOS and Privacy Policies described above. However, because Google had a separate contract with each educational institution, “Google [did] not mandate how these educational institutions receive[d] such authorizations, nor [was] that process uniform between various educational institutions.” Id. There were “substantial differences between how each of the [educational] institutions approaehe[d] disclosures,” and “it is unclear ... what disclosures ... each [GAFE] user saw before registering for an account.” Id.; see also id. at *15 (“Google had no single policy that required all Google Apps Administrators to provide the same disclosures to end users.”). “Some institutions’ disclosures [were] quite explicit. For example, Western Piedmont Community College [told] its users that ‘Google does use software or a “bot” to scan Gmail emails for key words for the purposes of targeted advertising.’ Similarly, the University of Aaska state[d] that ‘For use in targeted advertising on Google’s other sites, and if your email is not encrypted, software (not a person) does scan your mail and compile keywords for advertising.’ Meanwhile, other universities, such as the University of the Pacific, merely incorporate^] Google’s disclosures by citing to the TOS and Privacy Policies.” Id. (citations and alterations omitted). Moreover, “it is not clear that end users even had to look at these disclosures before they could create them accounts.” Id. As a final point, Google contends that, pursuant to certain contractual terms, many educational institutions chose to have “advertisements in GAFE ... off by default, meaning that advertisements were not displayed to GAFE users unless the GAFE institution affirmatively decided to allow them. Moreover, if a GAFE institution did not affirmatively enable advertisements, any automated scanning of GAFE emails was not used to serve targeted ads in GAFE Gmail or in any other Google service, nor were any ‘advertising profiles’ made.” ECF No. 66 at 4 (footnote omitted). In light of these circumstances, this Court concluded in Gmail that whether a GAFE user had consented to the interception of his or her communications — and thus whether Google had violated the Wiretap Act — was a question that “must be litigated on an individual, rather than elasswide basis.” 2014 WL 1102660, *14. Accordingly, plaintiffs in Gmail had failed to satisfy Federal Rule of Civil Procedure 23(b)(3), which states that a court may certify a damages class only if “the [281] court finds that the questions of law or fact common to class members predominate over any questions affecting only individual members.” Fed. R. Civ. P. 23(b)(3). Following the Court’s class certification ruling, only the individual plaintiffs’ individual claims remained. Google and the individual plaintiffs settled the individual claims and filed stipulations of dismissal as to all cases. The case was closed on July 14, 2014. B. Procedural History On January 27, 2016, Plaintiffs filed the original complaint in Corley. EOF No. 1. There were four Plaintiffs in the original complaint, all of whom were current or former students of the University of California, Berkeley (“Berkeley”). These Plaintiffs were Ryan Corley, William Dormann, Shannon Mehaffey, and Teddey Xiao. On February 8, 2016, Plaintiffs’ counsel filed a notice of voluntary dismissal as to Teddey Xiao. EOF No. 12. Corley was originally assigned to U.S. Magistrate Judge Nathanael Cousins, but was related to Gmail, and reassigned to the undersigned judge on February 12, 2016. On March 11, 2016, Plaintiffs’ counsel in Corley sought leave to file the FAC in order to add “an additional 700 plaintiffs,” EOF No. 18 at 2. On March 16, 2016, without obtaining leave from the Court, Plaintiffs’ counsel filed the FAC. Because this Court had not previously overseen a ease involving more than 700 Plaintiffs, the Court set an initial case management conference for April 20, 2016 to discuss with the parties “how best to manage this case” going forward. EOF No. 68 at 2. In the parties’ case management statement, Plaintiffs proposed a discovery plan in which Plaintiffs’ counsel wanted to proceed effectively as a class action: “Discovery should proceed as to ... a representative sample of the Plaintiffs ... [which] should help to value and settle any remaining cases.” EOF No. 66 at 13. At the case management conference itself, Plaintiffs’ counsel acknowledged that the Gmail class certification order influenced Plaintiffs’ decision to bring an action with 700 individual Plaintiffs. EOF No. 74-1 (“Hr’g Tr.”) at 5 (“There are issues raised by your [Gmail] order that are in play, yes.”). Further, when asked why Plaintiffs’ counsel did not file individual cases, Plaintiffs’ counsel stated that “[tjhey’re all the same essential claim arising out of the same essential facts.” Id. at 9. However, on consent, Plaintiffs’ counsel acknowledged that “[t]here may be differences based on school” in terms of what information each Plaintiff saw and what each Plaintiff knew about Google’s email scanning practices. Id. at 9. Plaintiffs’ counsel conceded that “consent may be individual in certain ways,” and there “may be some unique differences based on who read what [and] when that are totally individual.” Id. at 8, 9, Plaintiffs’ counsel thus acknowledged that the predominance issues in Gmail would also be an issue in Corley. In addition, Google argued that the FAC lacked sufficient information as to each Plaintiff. The FAC did not, for instance, give “individual e-mail addresses” for each Plaintiff. Id. at 34. Without this information, Google could not gather data or determine how to marshal a proper defense as to each Plaintiff. See id. at 15 (“There’s a lot of just basic biographical data that we would want to get our hands on sooner than later.”). During the case management conference, the Court commented that the FAC appeared to be “trying to circumvent Rule 23,” the Federal Rule of Civil Procedure governing class certification. Id. at 42. At the April 20, 2016 initial case management conference, Google stated that it had ceased scanning emails of GAFE users on April 30, 2014. Id. at 17. The Wiretap Act is subject to a two-year statute of limitations. Accordingly, Plaintiffs’ counsel stated, also at the initial case management conference, that he would file another ease which would allege a Wiretap Act claim against Google by several hundred more plaintiffs. Id. at 23 (Google’s counsel noting that he understood that Plaintiffs’ counsel would “add another 100 or 200 people to the case); id. at 32 (Plaintiffs’ counsel stating that he had 865 total clients, of which approximately 700 were named in the FAC). The Court responded that Plaintiffs [282] could not file such an action in federal court, because doing so would be “improper.” Hr’g Tr. at 33. Plaintiffs’ counsel then stated that he would “file it in Superior Court and presumably Google will remove it.” Id. at 32. Following the initial case management conference, the Court ordered the parties to file briefs addressing the propriety of mass join-der. ECF No. 69 at 1. The Court also scheduled a further case management conference for April 28, 2016. Id. Plaintiffs filed their brief on mass joinder on April 25, 2016, ECF No. 74, and Google filed its brief on April 26, 2016, ECF No. 77. At the April 28, 2016 case management conference, Plaintiffs’ counsel confirmed that he would file a separate action with more than a hundred Plaintiffs in state court. See also ECF No. 85 at 4 (Plaintiffs’ counsel stating that he would “file [the additional] claims separately in state court.”). After reviewing the parties’ briefs and learning about Plaintiffs’ counsel’s decision to file a separate action in state court, the Court invited Google at the April 28, 2016 case management conference to file a motion to sever. ECF No. 85. Plaintiffs’ subsequently filed a case with 175 Plaintiffs — Amaral—in Santa Clara County Superior Court on April 29, 2016. Google removed Amaral to federal court on May 11, 2016, and Amaral was related to Corley on May 18, 2016. 16-CV-2553, ECF Nos. 1 & 15. Google moved to sever on May 18, 2016 in Corley and on June 1, 2016 in Amaral. Google’s motion in Amaral states that “the Court should.. .sever the claims of the individual Plaintiffs in the present Amaral complaint” for “all the reasons set forth in the Corley Motion.” 16-CV-2553, ECF No, 20 at 2. There is no independent legal analysis in the Amaral motion. Plaintiffs filed responses to both motions on July 6, 2016, and Google filed its replies on July 27, 2016. ECF No. 121 (“Opp’n”); ECF No. 123 (“Reply”). On June 21, 2016, Public Citizen, Inc. (“Public Citizen”) moved to file an amicus brief opposing Google’s motion to sever. ECF No. 114. The Court granted Public Citizen’s motion on June 23, 2016, and Public Citizen filed its amicus brief on June 24, 2016. ECF No. 118 (“Public Citizen Opp’n”). The Court held a ease management conference on August 18, 2016 addressing options to move this litigation forward if the Court were to find mis-joinder. ECF No. 133. II. LEGAL STANDARD Under Federal Rule of Civil Procedure 20(a), permissive joinder of plaintiffs is appropriate where (1) “any right to relief [is asserted] jointly, severally, or in the alternative with respect to or arising out of the same transaction, occurrence, or series of transactions or occurrences,” and (2), “any question of law or fact common to all plaintiffs will arise in the action.” Fed. R. Civ. P. 20(a). “Even once these requirements are met, a district court must examine whether permissive joinder would comport with the principles of fundamental fairness or would result in prejudice to either side.” Coleman v. Quaker Oats Co., 232 F.3d 1271, 1296 (9th Cir.2000) (internal quotation marks omitted). “A determination on the question of joinder of parties lies within the discretion of the district court.” Wynn v. Nat. Broadcasting Co., 234 F.Supp.2d 1067, 1078 (C.D.Cal.2002); see also Acevedo v. Allsup’s Convenience Stores, Inc., 600 F.3d 516, 522 (5th Cir.2010) (per curiam) (“[District courts have considerable discretion to deny joinder when it would not facilitate judicial economy and when different witnesses and documentary proof would be required for plaintiffs’ claims.”). The rules regarding permissive joinder are “to be construed liberally in order to promote trial convenience and to expedite the final determination of disputes.” League to Save Lake Tahoe v. Tahoe Reg. Planning Agency, 558 F.2d 914, 917 (9th Cir.1977). III. DISCUSSION The Court begins by examining the first Rule 20(a) requirement — whether Plaintiffs’ right to relief arises out of the same transaction, occurrence, or series of transactions or occurrences.2 Because Plaintiffs have not met [283] this requirement, the Court need not consider whether Plaintiffs satisfy the second Rule 20(a) requirement: that there be a common question of law or fact. Aside from these two Rule 20(a) requirements, the Court also explains why mass joinder would be fundamentally unfair. Finally, the Court addresses how this litigation should move forward. A. Same Transaction or Occurrence Requirement 1. Legal Framework On the same transaction or occurrence requirement, the Ninth Circuit has held that “this provision requires factual similarity in the allegations supporting [plaintiffs’ claims.” Visendi v. Bank of Am., N.A., 733 F.3d 863, 870 (9th Cir.2013). Other courts have likewise determined that there must be a logical relationship between events, In re Prempro Prods. Liab. Litig., 691 F.3d 613, 622 (8th Cir.2010), and “that joinder is not appropriate where different products or processes are involved,” In re EMC Corp., 677 F.3d 1361, 1369 (Fed.Cir.2012). With this guidance in mind, the Ninth Circuit, in Visendi v. Bank of America, determined that joinder was improper where a group of 160 named plaintiffs sued 15 different defendants for illicit lending practices. 733 F.3d at 866. The Visendi court held that plaintiffs’ “loans [liad been] secured by separate properties scattered across the country, and some of the properties, but not all, [had been] sold in foreclosure.” Id. at 870. Thus, “[w]hile Plaintiffs allege in conclusory fashion that [defendants’ misconduct was ‘regular and systematic,’ them interactions with [defendants were not [in fact] uniform.” Id. Accordingly, the Visendi court determined that plaintiffs’ allegations did not arise from a single transaction or occurrence: they arose from different loans, made by different lenders, on different properties, in different parts of the country. Similarly, in Coughlin v. Rogers, 130 F.3d 1348 (9th Cir.1997), 49 plaintiffs brought suit against the Immigration and Naturalization Service (“INS”) and its then-Director, Richard Rogers. Id. at 1349. The gist of plaintiffs’ complaint was that “defendants have unreasonably delayed [the] adjudication] [of] plaintiffs’ applications and petitions.” Id. These applications and petitions, however, were significantly different from one another. Some petitions had been made by U.S. citizens, others by aliens. Some applications sought relief for the applicant, while others sought relief for a spouse or child. On these facts, the district court found that plaintiffs’ claims did not arise from the same transaction or occurrence: “[e]ach [p]laintiff has waited a different length of time, suffering a different duration of alleged delay.” Id. at 1350. “Furthermore, the delay is disputed in some instances and varies from case to case. And, most importantly, there may be numerous reasons for the alleged delay.” Id. Thus, “[e]ach” plaintiffs claim “raises potentially different issues, and must be viewed in a separate and individual light by the [c]ourt.” Id. at 1351. On appeal, the Ninth Circuit held that the district court’s ruling was consistent “with Federal Rules of Civil Procedure 20 and 21 and the precedent on severance.” Id. Although the Ninth Circuit has not addressed the propriety of joinder in a case involving as many as 879 Plaintiffs, several district courts in the Ninth Circuit have done so. In On the Cheap, LLC v. Does 1-5011, 280 F.R.D. 500 (N.D.Cal.2011), for example, the district court held that plaintiff had failed to satisfy the same transaction or occurrence requirement in bringing suit against 5000 Doe defendants over copyright infringement. As the court pointed out, defendants in On the Cheap would “likely raise different factual and legal defenses,” because the allegations made against them were all somewhat different. Id. “Because the large number of defendants with individual issues will create ‘scores of mini-trials involving different evidence and testimony and complicate the issues for all those involved, it is more efficient to proceed with separate cases where there will be separate proceedings, including separate motion hearings and ADR efforts.” Id. at 503. Other district courts are in accord with the reasoning in On the Cheap. See, e.g., WiAV Networks, LLC v. 3Com Corp., 2010 WL 3895047, *2 (N.D.Cal. Oct. 1, 2010) [284] (“Each defendant has simply been thrown into a mass pit with others to suit plaintiffs convenience. In this connection, the accused defendants — who will surely have competing interests and strategies — are also entitled to present individualized assaults on questions of non-infringement, invalidity, and claim construction”).3 In addition, consistent with On the Cheap, a number of circuit courts have found joinder of hundreds or thousands of parties inappropriate. In AF Holdings, LLC v. Does 1-1058, 752 F.3d 990 (D.C.Cir.2014), for example, the D.C. Circuit rejected plaintiffs’ attempt to join 1058 Doe defendants together for purposes of asserting a copyright infringement claim. As in On the Cheap, the crux of plaintiffs’ contention in AF Holdings was that plaintiffs owned a copyrighted work which had been downloaded numerous times by defendants through an online filesharing program. As the D.C. Circuit explained, two “users who download the same file months apart are like two individuals who play at the same blackjack table at different times. They may have won the same amount of money, employed the same strategy, and perhaps even played with the same dealer, but they have still engaged in entirely separate transactions. And simply committing the same type of violation in the same way does not link defendants together for the purposes of joinder.” Id. at 997-98 (internal quotation marks and alteration omitted). Looking outside the filesharing context, in Acevedo v. Allsup’s Convenience Stores, Inc., the Fifth Circuit determined that the district court properly exercised its discretion in denying joinder to 800 “current or former employees of Allsup’s Convenience Stores.” 600 F.3d at 518. These 800 employees sought to bring wage and hour claims against defendant. As the Fifth Circuit pointed out, however, “plaintiffs work or have worked across a network of more than 300 stores, each with its own manager responsible for implementing Allsup’s policies.” Id. at 522. “[Conditions do not appear to have been uniform at all Allsup’s stores.” Id. Thus, “given the divergent working condictions at each store and the different defenses applicable to each plaintiffs claims,” mass joinder was improper. Id.; see also id. (noting that joinder would be inefficient, as it would be “too challenging logistically”). Notably, the Fifth Circuit observed that plaintiffs in Acevedo had “fail[ed] to cite any cases in which a group of plaintiffs even remotely as numerous as 800 were able to join their claims.” Id. Two common themes link Visendi, Cough-lin, On the Cheap, AF Holdings, and Acevedo together. First, the single transaction or occurrence requirement is not met where plaintiffs would have to prove their claims or defendants would have to litigate their defenses on an individualized basis. Second, the impropriety of joinder is magnified in situations where there are hundreds or thousands of parties, as these situations burden, rather than facilitate, judicial economy. 2. Application These two principles govern the Court’s disposition of the instant motions. As noted in Gmail, “[t]he question of whether Class members have consented to the alleged interceptions has been central to this case since its inception.” 2014 WL 1102660, *13; see also id. (“The consent exception remains one of the principal disputed issues in this case.”). Consent is also central to the instant actions: Google has, since the initial ease management conference, stated that it would challenge whether Plaintiffs had consented to Google’s interception and scanning practices. Consent may be either express or implied, but both express and implied consent are [285] questions of fact. Id. at *15 (“[T]he question of express consent is usually a question of fact, where a fact-finder needs to interpret the express terms of any agreements to determine whether these agreements adequately notify individuals regarding the interceptions.”); id. at *16 (“Implied consent is an intensely factual question that requires consideration of the circumstances surrounding the interception to divine whether the party whose communication was intercepted was on notice that the communication would be intercepted.”). The fact-specific nature of the consent inquiry defeated class certification in Gmail, as “[t]he individualized questions with respect to consent, which will likely be Google’s principal affirmative defense, [were] likely to overwhelm any common issues.” Id. at *21. This inquiry also defeats mass joinder here. The “intensely individualized” nature of consent means that the 879 Plaintiffs’ individual claims do not satisfy the same transaction or occurrence requirement. The consent analysis instead requires an examination of many different transactions or occurrences. Indeed, consent is particularly fraught as it applies to GAFE users, because each educational institution had a unique contract with Google. This is a fact that Plaintiffs’ counsel acknowledged at the initial case management conference: “consent may be individual in certain ways,” and “there may be differences based on school, there may be differences based on time period, there may be some unique differences based on who read what [and] when that are totally individual.” Hr’g Tr. at 5-6. These differences are also readily apparent from the FAC in the instant cases: “Google did not mandate how these Educational Institutions obtained these authorizations, and the disclosures and authorizations regarding GAFE naturally varied by school.” FAC ¶ 20. The FAC further alleges that, “many Educational Institutions directed students to the school’s own privacy or use policies, some to Terms of use specifically agreed-upon by the Education Institution and Google, others to the Google Apps for Education Terms of Service and/or Privacy Policy, yet others to the general Google Terms of Service and/or Privacy Policies.” Id. ¶ 21. • The FAC describes the variations in the disclosures of some educational institutions. For example, in response to whether Google would “search our mail for marketing purposes,” the University of Maine (“Maine”) stated: “No. The education agreement prohibits that activity.” Id. ¶ 23(7). The University of Arizona (“UA”) stated that “CatMail allows you to send and receive email with all the Google tools ... plus full privacy of your information and no ads.” Id. ¶ 23(6). It is unclear whether “full privacy” included privacy from email scanning. Similarly, Google’s contract with Yale University (“Yale”) “guarantee[d] privacy and confidentiality in ways that are analogous to the way Yale currently provides email services,” but did not identify what privacy protections Yale’s current email service provided. Id. ¶ 23(5). Aside from institutions like Maine, UA, and Yale, the FAC also states that some educational institutions simply “directed students to ... the general Google Terms of Service and/or Privacy Policy.” FAC ¶ 21. Those TOS and Privacy Policies, however, changed multiple times between 2010 and 2014. On March 1, 2012, Google amended its TOS such that all Gmail users would now have to agree to “give Google ... a worldwide license to use ... and distribute [user] content.” Id. at *3. 2014 WL 1102660, *3. That same day, the Privacy Policy was also changed to state “that Google could collect information that users provided to Google, cookies, log information, user communications to Google, information that users provide to affiliated sites, and the links that a user follows.” Id.4 The Privacy Policy was changed again in April 2014, after the Court’s Gmail class certification order, to inform users that “[o]ur automated systems analyze your content (including emails) to provide you personally relevant product fea [286] tures, such as customized search results, tailored advertising, and spam and malware detection.” 15-CV-4062, EOF No. 20-2 at 3. Thus, any analysis of consent would have to address the specific TOS and Privacy Policies in effect at specific times. Finally, according to Google, many educational institutions agreed, as a default, that GAFE users would not be subject to targeted advertising and that Google would not scan email sent or received by a GAFE user. See ECF No. 66 at 4 (“[I]f a GAFE institution did not affirmatively enable advertisements, any automated scanning of GAFE emails was not used to serve targeted ads in GAFE Gmail or in any other Google service, nor were any ‘advertising profiles’ made.”). In order to enable scanning, the educational institutions themselves or GAFE users at these educational institutions had to affirmatively change these default settings. This dissimilarity reflects the individual contracts that each educational institution had with Google. See FAC ¶ 23(12) (“NYU Google Apps for education is offered by Google to NYU in accordance with a specially negotiated end-user license agreement.”) (emphasis added).These differences impacted what Google scanned and analyzed, what disclosures each Plaintiff read, and whether any particular Plaintiff consented to Google’s practices. As this Court has explained in Gmail, the “consent of Education Class members is likely to require individualized inquiries .... Google had no single policy that required all Google Apps Administrators to provide the same disclosures to end users. This means that the end users received vastly different disclosures depending on with which educational institution they were affiliated.” 2014 WL 1102660, *16. Put another way, each school had its own contract with Google, and these contracts appear to have provided different disclosures. Based on this information, Plaintiffs at each educational institution (1) might not have consented to Google’s scanning practices, (2) might have opted out of Google’s scanning practices by changing their Gmail settings or choosing a different email provider, or (3) might have consented to Google’s scanning practices. As Google notes, in order to defend itself, Google would need to undertake “a highly individualized” analysis to investigate, research, and present defenses as to each Plaintiff. Mot. at 10. This type of investigation made class treatment inappropriate. It also makes mass joinder inappropriate. For the 879 Plaintiffs, spread across twenty one educational institutions, Google would need to answer a set of different questions. These questions might include: • Did a particular Plaintiffs educational institution have a specific privacy policy, or did it direct users to Google’s general TOS and Privacy Policies? • What did each institution’s privacy policy disclose or authorize? • Did the particular Plaintiff actually read the policies at issue? • Did an educational institution’s privacy policies change over time? Which version or versions of an educational institution’s privacy policy did the particular Plaintiff read? • Did an educational institution allow its users to change GAFE settings to prevent or enable email scanning? ® Did the particular Plaintiff hear about Google’s interception practices from any another source, such as a Gmail-related privacy notice or a publiely-available news article? Courts have recognized that the predominance analysis under Rule 23(b) and the single transaction or occurrence analysis under Rule 20(a) may, in some instances, be two sides of the same coin. In In re EMC Corp., for instance, defendants sought to sever themselves from eighteen other companies that were named in a single complaint. 677 F.3d at 1363. In evaluating defendants’ arguments, the Federal Circuit examined an “analogous” Fifth Circuit decision where the Fifth Circuit had determined that plaintiffs had “not established that the questions of law or fact common to the members of the class predominate over any questions affecting individual members as required by Rule 23(b)(3).” Id. at 1355. The Federal Circuit acknowledged that “[t]he procedural safeguards pertaining to class certification do not [technically] apply” in a joinder case. Id. [287] Nevertheless, as the Federal Circuit explained, “Rule 20’s requirements are designed to prevent similar unfairness.” Id. “Like the requirements of Rule 23 in class actions, Rule 20’s two requirements — that the claims share questions of law or fact common to all defendants, and arise out of the same transaction or occurrence — help ensure that the scope of the action remains consistent with fairness to the parties.” Id. (internal quotation marks and alterations omitted). Here, the reasoning in In re EMC is particularly relevant because Plaintiffs’ counsel has acknowledged that the Corley and Amaral cases are “much like ... class action[s].” ECF No. 74 at 7. Plaintiffs’ counsel, moreover, has acknowledged that the Corley and Amaral cases were brought to avoid the reach of Gmail, In opposing the instant motions, for instance, Plaintiffs’ counsel has stated that “Plaintiffs seek to litigate their claims ... consistent with the law, in light of the unavailability of class treatment.” Opp’n at 1 (emphasis added). Moreover, Plaintiffs’ counsel has stated that discovery should proceed in a manner that mimics discovery in a class action: “[discovery should ... focus on a representative sample of Plaintiffs,” which would culminate in a limited set of “bellwether trials.” ECF No. 131 at 4. Plaintiffs concede that “[djeposing all 879 Plaintiffs before trying the first case is a waste of resources for the Parties and the Court.” Id. Given these circumstances, the Court’s Rule 23(b)(3) predominance analysis and Rule 20(a) same transaction or occurrence analysis reach the same result. As before in Gmail, Google must have the opportunity to fully and fairly litigate against each Plaintiff in this ease. That opportunity allows Google to review and present defenses against each Plaintiff, given the individualized nature of the consent exemption. Thus, the Wiretap Act claims of the 879 Plaintiffs do not arise from the same transaction or occurrence, and Plaintiffs have been misjoined. 3. Plaintiffs’ and Public Citizen’s Contentions Plaintiffs and Public Citizen present three arguments in response to this conclusion. First, they point out that some courts have denied class certification but have nonetheless found joinder to be appropriate. Second, Plaintiffs and Public Citizen argue that Rule 20(a) is satisfied because “Google had a uniform policy of intercepting and processing the content of ... email[ ].” Public Citizen Opp’n at 7. Third, Plaintiffs and Public Citizen cite the Court’s prior rulings in the McAfee v. Francis litigation. As discussed below, these arguments lack merit. a. Predominance and Joinder According to Public Citizen, the Court’s predominance analysis in Gmail does not necessarily foreclose joinder here. To support this point, Public Citizen cites several cases where courts “have allowed joinder under Rule 20 after denying class certification under Rule 23.” Public Citizen Opp’n at 5. In Bryant v. Service Corp. International, 801 F.Supp.2d 898, 900 (N.D.Cal.2011), the court denied class certification to a putative class of current and former Service Corp. employees. There were twelve plaintiffs in Bryant. After the court denied class certification, the court severed the eight plaintiffs who did not reside in California. The district court, in its discretion, declined to sever the four California plaintiffs. Similarly, in Rochlin v. Cincinnati Insurance Co., 2003 WL 21852341, *3-*4 (S.D.Ind. July 8, 2003), Weigele v. FedEx Ground Package System, Inc., 2010 WL 3069213, *1 (S.D.Cal. Aug. 4, 2010), and Luna v. Del Monte Fresh Produce (Southeast), Inc., 2009 WL 4801357, *1 (N.D.Ga. Dec. 10, 2009)), the courts agreed to join 9, 25, and 52 employees of a single employer, respectively. None of the foregoing eases involved an attempt to join 879 Plaintiffs at 21 different educational institutions together. The largest number any court permitted in terms of join-der was 52, in Luna. Furthermore, these cases involve employees suing the same employer for wage, hour, or other labor violations. In such circumstances, the parties may well be able “to take advantage of the efficiencies of common discovery and joint proceedings” — there might have, for instance, been a single employment contract or collective bargaining agreement. Weigele, 2010 [288] WL 3069213, *2. In the instant actions, on the other hand, there are no such economies of scale. There are 21 contracts, one for each educational institution. There is uncertainty as to what each Plaintiff saw, whether a Plaintiff agreed to a particular disclosure, and whether a Plaintiff might have learned about Google’s practices from other sources, such as media coverage. Plaintiffs’ counsel conceded this point and stated at the initial case management conference that “consent may be individual in certain ways,” and there “may be some unique differences based on who read what [and] when that are totally individual.” Hr’g Tr. at 8, 9. Based upon these differences, Google intends to examine each Plaintiffs Wiretap Act claim on an individualized basis in order to determine whether a Plaintiff expressly or impliedly consented to Google’s practices. The employer-employee cases cited by Public Citizen are therefore inapposite, and the Court declines to follow them here. b. Uniform Policy Next, Plaintiffs and Public Citizen claim that Google had a uniform policy of intercepting and scanning email. However, Google contends that, “[s]ince GAFE was launched, advertisements in GAFE Gmail have been off by default, meaning that advertisements were not displayed to GAFE users unless the GAFE institution affirmatively decided to allow them. Moreover, if a GAFE institution did not affirmatively enable advertisements, any automated scanning of GAFE emails was not used to serve targeted ads in GAFE Gmail or in any other Google service, nor were any ‘advertising profiles’ made.” ECF No. 66 at 4 (footnote omitted). Thus, different educational institutions may have had different scanning policies. Furthermore, even if Google had a uniform scanning policy across all educational institutions, each educational institution had its own privacy policy that included different authorizations and disclosures. The instant actions are therefore distinguishable from three decisions cited by Plaintiffs and Public Citizen: United States v. Mississippi, 380 U.S. 128, 85 S.Ct. 808, 13 L.Ed.2d 717 (1965), Nelson-Devlin v. Eli Lilly and Co., 2015 WL 5436700 (E.D.Cal. Sept. 15, 2015), and Allen v. Similasan Corp., 2013 WL 2120825 (S.D.Cal. May 14, 2013). In Mississippi, plaintiff alleged that Mississippi had implemented a systemic, uniform plan to restrict the African American vote, implemented via certain statewide voting procedures. The U.S. Supreme Court held that joinder of six county registrars was proper because “the registrars had acted and were continuing to act as part of a state-wide system designed to enforce the registration laws in a way that would... deprive colored people of the right to vote.” 380 U.S. at 142, 85 S.Ct. 808. Similarly, in Nelson-Devlin, the district court found joinder of plaintiffs proper because the warnings accompanying prescriptions of a specific drug were found to be identical. Likewise, in Allen, the plaintiffs had purchased the same product after reviewing the same “product[ ] packaging” and the same advertisements on a website. 2013 WL 2120825, *2. Unlike this trio of cases, the instant cases involve potentially different scanning policies across different educational institutions as well as different authorizations and disclosures. Consequently, legal authority upholding joinder in Mississippi, Nelson-Devlin, and Allen, which involved a uniform policy, is inapplicable here. c. McAfee Litigation Finally, Plaintiffs and Public Citizen make much of this Court’s decisions in McAfee v. Francis. In McAfee v. Francis (“McAfee I"), 2011 WL 3293759 (N.D.Cal. Aug 1, 2011), three plaintiffs originally brought breach of contract claims against a single defendant. The Court granted defendant’s motion to dismiss without prejudice because plaintiffs did not contest defendant’s arguments on the merits; plaintiffs simply requested leave to amend. In dicta, the Court observed that joinder appeared improper because it appeared that the plaintiffs had separate contracts with defendant. Id. at *3. Plaintiffs thereafter filed an amended complaint, and defendant once again moved to dismiss. Notably, in the amended complaint, plaintiffs asserted claims for breach of contract, negligent misrepresentation, and fraud. The Court granted in part and denied in part [289] defendant’s second motion to dismiss in McAfee v. Francis (“McAfee II”), 2012 WL 762118 (N.D.Cal. Mar. 6, 2012). On joinder, the Court held that plaintiffs’ breach of contract claims were “predicated on different transactions and occurrences,” as each plaintiff had a separate contract with each defendant. “However, some courts have concluded that even when the underlying contract claims are predicated on different transactions and occurrences, allegations of fraud and misrepresentation may constitute a transaction or series of transactions under Rule 20(a) when they are part of a common scheme.” Id. The Court agreed with this approach, and held that the amended complaint’s allegations of a common fraudulent scheme were sufficient to constitute a “logical relationship” for purposes of Rule 20(a). Id. Unlike McAfee, violations of the Wiretap Act do not sound in fraud. As the Third Circuit has explained, “[t]he Wiretap Act is a wiretapping statute, and just because a scenario sounds in fraud or deceit does not mean it sounds in wiretapping.” In re Google Inc. Cookie Placement Cons. Privacy Litig., 806 F.3d 125, 144 (3d Cir.2015). Neither Plaintiffs nor Public Citizen cites any case where courts have applied the “fraud” or “misrepresentation” rule employed in McAfee II to the Wiretap Act, and the Court has found none in its own research. To summarize, because Plaintiffs have failed to meet the same transaction or occurrence requirement, the Court finds that Plaintiffs have been misjoined in the Corley and Amaral cases. Because Plaintiffs have not met the same transaction or occurrence requirement, the Court need not address whether Plaintiffs have satisfied Rule 20(a)’s second requirement — that there be a common question of law or fact.5 B. Fundamental Fairness As noted above, the Court may, in the interest of fundamental fairness and justice, dismiss or sever parties even if the Rule 20(a) requirements are satisfied. Here, the Court finds that, if the Corley and Amaral actions were to proceed as is, joinder (1) would cause prejudice to Google and result in jury confusion, (2) would be impractical, and (3) would avoid proper payment of fees. For these reasons, the Court finds that dismissal or severance would better comport with principles of fundamental fairness and would better advance the administration of justice. 1. Prejudice to Google and Jury Confusion First, keeping 879 Plaintiffs in these actions would cause significant prejudice to Google. Indeed, several courts have found substantial prejudice where a “defendant’s defense would, in effect, require [hundreds of] mini-trial[s].” Hard Drive Prods., Inc. v. Does 1-188, 809 F.Supp.2d 1150, 1164 (N.D.Cal.2011). In addition, in evaluating the likelihood of prejudice, the Ninth Circuit has advised courts to consider the “danger of jury confusion.” Coleman, 232 F.3d at 1296. Thus, in Coleman, the Ninth Circuit determined that certain plaintiffs were misjoined where all ten plaintiffs had alleged age discrimination against the same employer. Id. As the Ninth Circuit explained, defendant “would be prejudiced by having all ten plaintiffs testify in [a single] trial” because “even the strongest jury instructions [can] not ... dull[ ] the impact of a parade of witnesses, each recounting his contention that defendant laid him off because of his age.” Id. Both concerns — the potential for hundreds of mini-trials and the likelihood of jury confusion — are at play here. As has been discussed, if all 879 Plaintiffs were to remain joined, Google would potentially have to take discovery of each Plaintiff, investigate the disclosures made by each educational institution, and determine whether to file hundreds if not thousands of pretrial motions. All of [290] these obstacles have been magnified even more by the deficiencies in the FAC. As Google’s counsel has noted, the FAC does not provide Plaintiffs’ individual e-mail addresses. Hr’g Tr. at 31. It is therefore difficult, if not impossible, for Google to “try to answer” Plaintiffs’ claims, Id. In moving to sever, Google reiterates that “the sheer volume of Plaintiffs and GAFE institutions, each with different fact patterns requiring review for liability and damages, would prejudice Google .., if Plaintiffs’ claims were tried together,” Mot. at 10. Moreover, on jury confusion, the Court notes that if this case were to go to trial, and if Google were to argue the consent exception, the jury would potentially have to keep track of the facts and circumstances of 879 different Plaintiffs. Moreover, the jury would have to consider and analyze different privacy policies to reflect the diversity of disclosures in this litigation. See Wynn, 234 F.Supp.2d at 1089 (“A single trial would present the jury with the hopeless task of trying to discern who did and said what to whom and for what reason.”). Plaintiffs do not present a persuasive counterargument to this point. In fact, Plaintiffs simply contend that few cases go to trial, and that the Court need not consider jury confusion at this stage of litigation. Plaintiffs’ arguments fail to comport with Ninth Circuit precedent, which expressly instructs courts to consider the risk of jury confusion in evaluating prejudice to the parties when analyzing misjoinder. See Coleman, 232 F.3d at 1296 (stating that court should examine “danger of jury confusion” in misjoinder analysis). Further, any additional delay on the issue of joinder would only cause the Court and the parties more difficulty in the future, as this action moves through fact and expert discovery, pretrial proceedings, and trial. Accordingly, contrary to Plaintiffs’ contentions, the Court finds that jury confusion and unfair prejudice to Google would result if all 879 Plaintiffs remained in these two actions. 2. Impracticality of Joinder Second, it would be impractical to litigate all 879 Plaintiffs’ claims together. The sheer number of Plaintiffs, combined with their geographic diversity, would impede effective case management and would not promote judicial economy. As to the sheer number of Plaintiffs, the Court finds instructive the Bridgeport Music, Inc. v. 11C Music, 202 F.R.D. 229 (M.D.Tenn.2001), decision. In Bridgeport, the court granted defendants’ motion to sever 770 plaintiffs. In so doing, the court noted that, “[a]s a practical matter, this ease is unmanageable in its current form.” Id. at 232. “Because this Court’s courtroom would seat only a small fraction of [defendants and them attorneys, it cannot even hold a hearing on the motions currently pending; it cannot host a management conference; it certainly cannot try all — or even most — of the [plaintiffs’ counts together.” Id.;see also Hard Drive Prods., 809 F.Supp.2d at 1164 (finding misjoinder based on presence of 188 defendants). The instant action, which involves 879 Plaintiffs, would present even more logistical hurdles than in Bridgeport Plaintiffs’ counsel, indeed, has already admitted that it would be both impractical and inefficient to require that all 879 Plaintiffs be deposed. Plaintiffs’ counsel, moreover, has provided no plan as to what the Court should do if all 879 Plaintiffs were required to attend a pretrial hearing or if all (or even a fraction of the) 879 Plaintiffs elected to attend trial. Case law interpreting Federal Rule of Civil Procedure 23(a)(1), commonly known as the numerosity requirement, provides yet another indicator as to the impracticality of join-der. This requirement states that “[o]ne or more members of a class may sue or be sued as representative parties on behalf of all members only if the class is so numerous that joinder of all members is impracticable.” Fed. R. Civ. P. 23(a)(1). There is no “ ‘magic number’ that will satisfy the Rule 23(a)(1) prerequisite in every case.” Newberg on Class Actions § 3:12 (6th ed. 2016). “As a general guideline, however,” a class “of 40 or more members raises a presumption of impracticability of joinder.” Id; see, e.g., Pa. Pub. Sch. Emps. Ret. Sys. v. Morgan Stanley & Co., Inc., 772 F.3d 111, 120 (2d Cir.2014) (“Numerosity is presumed for classes larger than forty members.”); Hernandez v. Cnty. [291] of Monterey, 305 F.R.D. 132, 153 (N.D.Cal.2015) (same); Huynh v. Harasz, 2015 WL 7015567, *5 (N.D.Cal. Nov. 12, 2015) (same). Having 879 Plaintiffs spread across two case dockets far exceeds this general guideline, particularly given the individualized inquiries necessary in the instant cases. Aside from the total number of Plaintiffs, the Court also observes that Plaintiffs reside in states across the country, from Massachusetts to Ohio to California. FAC ¶ 17(1), 17(6), 17(450). Numerous courts have held that, “[i]n assessing impractieality of join-der,” the court should “consider factors such as the geographical diversity of [parties].” McCluskey v. Trs. of Red Dot Corp. Emp. Stock Ownership Plan and Trust, 268 F.R.D. 670, 674 (W.D.Wash.2010) (internal quotation marks omitted); accord Pole v. Estenson Logistics, LLC, 2016 WL 4238635, *5 (C.D.Cal. Aug. 10, 2016) (same). Geographic diversity may make it more difficult for the Court to hold hearings, for the parties to conduct discovery, and for the parties to attend trial and pretrial proceedings. Hence, the geographic diversity of Plaintiffs weighs against mass joinder. The logistical hurdles described above are not mere hypotheticals. After Plaintiffs filed the FAC in Corley — without the Court’s authorization — these 700 different names had to be inputted into the Court’s Electronic Case Filing (“EOF”) system. This input took place over two days, March 16 and 17, 2016. Over 700 names were again inputted into the EOF system after Plaintiffs filed their proposed protective order on May 13, 2016. These burdens will continue through discovery, pretrial proceedings and trial. Hard Drive Prods., 809 F.Supp.2d at 1164 (declining to allow joinder because joinder “would result in a logistically unmanageable case” and would not “promot[e] judicial economy and trial convenience”). In an attempt to counter these impracticability concerns, Public Citizen cites several cases where courts have “entertained suits with hundreds of plaintiffs joined together in one action under Rule 20.” Public Citizen Opp’n at 3. None of these decisions, however, even address the propriety of joinder. Instead, all of these decisions focused on whether federal subject matter jurisdiction was proper. Five cases involved motions to remand under the Class Action Fairness Act (“CAFA”), and the remaining case involved a motion to remand under the Price Anderson Act (“PCA”). Neither Plaintiffs nor Google, on the other hand, have filed a motion to remand or challenged whether federal subject matter jurisdiction is proper. In two of the cases cited by Public Citizen, the court found that remand was appropriate. As such, the matter could not remain in federal court, and the court could not consider joinder issues. See Armstead v. Multi-Chem Gp., LLC., 2012 WL 1866862, *1 (W.D.La. May 21, 2012) (granting motion to remand); Aburto v. Midland Credit Mgmt., Inc., 2009 WL 4884147, *1 (N.D.Tex. Dec. 16, 2009) (same). In three of the remaining four decisions, the court did not even mention Rule 20(a), much less engage in any sort of discussion on joinder. See Bullard v. Burlington N. Santa Fe Ry. Co., 535 F.3d 759 (7th Cir.2008) (finding CAFA jurisdiction); Acuna v. Brown & Root Co., 200 F.3d 335 (5th Cir.2000) (finding jurisdiction under PCA); Adams v. Macon Cnty. Greyhound Park, Inc., 2011 WL 5513210 (M.D.Ala. Nov. 10, 2011) (finding CAFA jurisdiction). Lastly, in Gilmore v. Bayer Corp., 2009 WL 4789406, *3 (S.D.Ill. Dec. 9, 2009), a putative class action case involving 100 plaintiffs, the distinct court limited its Rule 20(a) discussion to whether plaintiffs had presented common questions of law and fact as to the side effects of a particular drug. The Gilmore court did not determine whether plaintiffs’ claims arose from the same transaction or occurrence, or whether joinder was fundamentally fair. Moreover, the Gilmore court focused not on Rule 20(a) — its joinder discussion was a mere three sentences long — but on whether plaintiffs had sufficiently met CAFA’s amount-in-eontroversy and minimum diversity requirements. Id. Accordingly, none of the cases Public Citizen cites supports a finding of joinder in the present case. These cases address federal subject matter jurisdiction, an issue that has [292] not been contested here. None of these cases involve an instance where class certification was previously denied, as it has been in the instant cases. In fact, all of the CAFA cases were putative class actions. After convening two hearings, examining two rounds of briefing, reviewing Public Citizen’s amicus brief, and conducting its own research, the Court reaches the same conclusion as that of the Fifth Circuit in Acevedo: the Court has failed to identify “any cases in which a group of plaintiffs even remotely as numerous as 800 were able to join their claims.” 600 F.3d at 622. 3. Avoidance of Fees Finally, several courts have found misjoin-der where a party’s actions appear to be motivated by an attempt to avoid filing fees. In CineTel Films, Inc. v. Does 1-1052, 863 F.Supp.2d 546, 664 n. 4 (D.Md.2012), for instance, the district court noted that the judicial burden in mass joinder cases is “compounded by the fact that the increased work resulting from mass joinder requires no additional payment beyond the one-time ... filing fee. Plaintiffs therefore in no way compensate financially for [their] significant drain on judicial resources.” Placing such an “enormous burden” on the parties and the court does not promote “the administration of justice.” Id. at 564. Likewise, inIO Group, Inc. v. Does 1-135, 2011 WL 445043, *6 n. 5 (N.D.Cal. Feb. 3, 2011), the district court noted that mass join-der, if left unchecked, would effectively “encourage [plaintiffs ... to join (or misjoin) as many doe defendants as possible,” all while allowing plaintiffs to dodge “filing fees.” Thus, “filing one mass action in order to identify hundreds of doe defendants through pre-service discovery and [to then] facilitate mass settlement,” all through a single filing fee, does not promote the purposes underlying joinder. Id. at *6. Here, Plaintiffs’ counsel has admitted that he chose to join all 879 Plaintiffs because “requiring Plaintiffs to pay $282,800 in filing fees and to file and litigate more than 700 + virtually identical individual actions [would] amount[ ] to a disproportionate and unwarranted burden.” Opp’n at 13. While Plaintiffs’ counsel might have relieved himself of “a disproportionate and unwarranted burden,” he has imposed “a disproportionate and unwarranted burden” on everyone else. As discussed above, Google faces an increased burden in terms of discovery, analysis of the consent exemption, and preparation for trial. The Court must also address how to move these cases forward in a fair, judicious, and efficient manner — a unique challenge given the 879 Plaintiffs with individual claims. Moreover, if Plaintiffs in fact lacked the financial resources to file a new set of complaints, Plaintiffs could avail themselves of the same opportunity offered to other litigants: an application to proceed in forma pauperis. Under the federal in forma pauper-is statute, a court may authorize the commencement of a suit without prepayment of the filing fee required by the clerk of the court if the plaintiff submits an affidavit of poverty showing that he or she is “unable to pay such fees or give security therefor.” 28 U.S.C. § 1915(a); see also Civil L.R. 3-10 (providing in forma pauperis procedures for Northern District of California). The Court raised this point at the initial case management conference on April 20, 2016. In the four months since, Plaintiffs’ counsel has yet to address why the in forma pauperis procedure would not alleviate their financial burden. Accordingly, consistent with the reasoning in CineTel and IO Group, the Court finds that Plaintiffs’ counsel’s attempt to avoid paying filing fees also weighs in Google’s favor. To conclude, if the Corley and Amaral cases were to proceed as is, there is a significant risk of jury confusion and prejudice to Google. Mass joinder would also be impractical, and would allow Plaintiffs to impose an unjustified burden on judicial resources. Given these conditions, mass joinder does not comport with fundamental fairness, and the Court finds that either dismissal or severance is appropriate. C. Case Management Having determined that joinder is improper, the next issue is how to move this litigation forward. This issue implicates two separate but related questions. First, should [293] Plaintiffs be severed, or dismissed without prejudice? Second, should Plaintiffs be required to proceed individually, through individual, separate complaints, or allowed to proceed by educational institution or some other grouping? The Court addresses these questions in turn. 1. Severance vs. Dismissal Federal Rule of Civil Procedure 21 provides that “[m]isjoinder of parties is not a ground for dismissing an action. On motion or on its own, the court may at any time, on just terms, add or drop a party.” Fed. R. Civ. P. 21. Alternatively, “[t]he court may.. .sever any claim against a party.” Id. “The effect of each option is quite different. When a court ‘drops’ a [party] under Rule 21, that [party] is dismissed from the ease without prejudice.” DirecTV, Inc. v. Leto, 467 F.3d 842, 845 (3d Cir.2006). “When that occurs, the statute of limitations is not tolled because we treat the initial complaint as if it never existed.” Id. (internal quotation marks omitted). “But when a court ‘severs’ a claim against a [party] under Rule 21, the suit simply continues against the severed [party] in another guise.” Id. “The statute of limitations is held in abeyance, and the severed suit can proceed so long as it initially was filed within the limitations period.” Id. “Because a district court’s decision to remedy misjoinder by dropping and dismissing a party, rather than severing the relevant claim [or party], may have important and potentially adverse statute-of-limitations consequences, the discretion delegated to the trial judge to dismiss under Rule 21 is restricted to what is ‘just.’” Id. “[District courts who dismiss rather than sever must conduct a prejudice analysis, including loss of otherwise timely claims if new suits are blocked by statutes of limitations.” Rush v. Sport Chalet, Inc., 779 F.3d 973, 975 (9th Cir.2015) (internal quotation marks omitted). In the present cases, the severance or dismissal decision does implicate the statute of limitations. Google ceased scanning GAFE emails on April 30, 2014. The Wiretap Act has a two-year statute of limitations. Both the Corley case, filed on January 27, 2016, and the Amaral ease, filed on April 29, 2016, fall within this statute of limitations, which would fall on April 30, 2016. However, if the Court were to dismiss all but the original three Plaintiffs in Corley, the statute of limitations would likely bar the other 876 Plaintiffs from bringing suit. At least some district courts in the Ninth Circuit have concluded that “any prejudice resulting from the statute of limitations is the risk assumed by the parties who caused the misjoinder.” Robinson v. Geithner, 2011 WL 66158, *9 (E.D.Cal. Jan. 10, 2011); accord Funtanilla v. Tristan, 2010 WL 1267133, *6 (E.D.Cal. Mar. 30, 2010) (“[A]ny prejudice resulting from the statute of limitations is a result of [p]laintiff combining seven unrelated claims in one complaint, instead of filing seven separate actions.”). Here, there is some evidence that Plaintiffs’ counsel assumed the risk that the Court would find misjoinder. At the April 20, 2016 initial case management conference, when Plaintiffs’ counsel stated his intention to file another ease with several hundred Plaintiffs, each with individual claims, the Court informed counsel that he could not do so because doing so would be “improper.” Hr’g Tr. at 31. Plaintiffs’ counsel then stated that he would “file it in Superior Court and presumably Google will remove it.” Id. at 32. After the April 20, 2016 initial case management conference, Plaintiffs and Google filed on April 25, 2016 and April 26, 2016, respectively, briefs addressing the propriety of mass joinder. Plaintiffs’ counsel confirmed at the April 28, 2016 case management conference that he would file another ease with more than a hundred Plaintiffs in state court. ECF No. 85 at 4. That case, filed on April 29, 2016 in Santa Clara County Superior Court, was Amaral. Thus, prior to April 30, 2016, the Wiretap Act’s statute of limitations deadline, Plaintiffs’ counsel (1) learned the relevant law regarding joinder in order to prepare his April 25, 2016 brief, and (2) knew that the complaints in Corley and Amaral may have been improper. Nevertheless, Plaintiffs’ counsel proceeded to file Amaral in state court on April 29, 2016, and decided not to [294] separate Plaintiffs in Corley into separate complaints. Two considerations, however, weigh against dismissal. First, the instant actions present several novel legal questions. No other court has addressed mass joinder in the Wiretap Act context. Public Citizen’s amieus brief further highlights the fact that the instant motions raise novel legal issues. Second, although the Ninth Circuit has not squai’ely addressed the issue, other circuits have generally held that a district court should sever rather than dismiss parties when the statute of limitations comes into play. In Elmore v. Henderson, 227 F.3d 1009, 1012 (7th Cir.2000), for instance, the Seventh Circuit stated that “[t]he district judge could and should have allowed [plaintiffs] claim against [defendant] to continue as a separate suit so that it would not be time-barred.” The Elmore court cited case law from the Third Circuit to support its conclusion. Consequently, given the fact that the instant case presents novel legal issues and that other courts generally caution against dismissal, the Court declines to dismiss 876 Plaintiffs. Instead, Plaintiffs shall be severed, pursuant to the approach outlined below. 2. Proceeding Through Individual Complaints Plaintiffs, if they elect to continue with this litigation, should prosecute their claims by individual Plaintiff. This means that the three original Plaintiffs — Ryan Corley, William Dormann, and Shannon Mehaffey — would remain in one action, but that the other 876 Plaintiffs would be severed, and Plaintiffs’ counsel would need to file 876 additional complaints. Each complaint must include specific factual allegations, which at a minimum would include each Plaintiffs GAFE email address, when each Plaintiff signed up for GAFE, what each Plaintiff saw and read in deciding to sign up for GAFE, and whether each Plaintiff ever changed his GAFE settings to allow or prevent Google’s interception and scanning. The Court reaches this conclusion based on the following six considerations. First, as outlined above, Plaintiffs have failed to meet the same transaction or occurrence requirement. Pursuant to the Ninth Circuit’s ruling in Visendi, there is a lack of factual similarity amongst the joined Plaintiffs. 733 F.3d at 870. Each Plaintiffs case rests upon an individualized consent analysis. Did a Plaintiff read a particular policy? Did those policies change over time? Did a Plaintiff change his or her privacy settings? Did a Plaintiff hear about Google’s practices through a third-party media source? These questions must be examined on a Plaintiff by Plaintiff basis. Second, the Court has considered severance by educational institution — an option advocated by Plaintiffs and Public Counsel— but finds that this option does not meaningfully address the same transaction or occurrence requirement or the fairness concerns discussed in this Order. Of the 879 Plaintiffs, 243 are affiliated with the University of California, Santa Cruz; 90 with the University of Arizona; 83 with the University of Michigan; 77 with New York University; 69 with the University of California, Berkeley; and 66 with New York University. ECF No. 129 at 3. An action involving 243, 90, 83, 77, 69, or 66 individual claims would be unmanageable given the individualized inquiries necessary in the instant cases. Third, and relatedly, severance by school would likely result in even more motions to sever. Google has in fact stated that “school-based actions would still present the individualized questions of whether each Plaintiff. .. consented to Google’s automated processing.” Reply at 14. The Corley case was originally filed on January 27, 2016. In the seven months since, the Court has, for purposes of addressing the issue of joinder, held three case management conferences focused on joinder, reviewed multiple rounds of briefing, and received an amicus brief. If the Court were to sever by educational institution, and if Google were to file another set of motions to sever, it would take months before the Court could tuna to the fundamental merits of Plaintiffs’ claim. That would dis-serve the parties, the Court, and the general public. Fourth, the Court has provided Plaintiffs’ counsel ample opportunity to address poten [295] tial issues involving misjoinder. Plaintiffs’ counsel was, as an initial matter, never even authorized to file the Corley FAC, which added over 700 Plaintiffs. Next, at the April 20, 2016 case management conference, the Court stated that Plaintiffs’ counsel’s filing of another case with several hundred Plaintiffs, each with individual claims, would be improper given the numerous unresolved issues concerning joinder in Corley. After the April 20 and 28, 2016 case management conferences, Plaintiffs’ counsel filed Amaral, a case in state court with 175 Plaintiffs on April 29, 2016, even though Plaintiffs’ counsel knew the case would be removed to federal court and related to the Corley ease before the undersigned judge. Fifth, such a decision is consistent with the reasoning in Gmail. In Gmail, the Court examined the differences between the disclosures provided at each university. The Court, for instance, compared the privacy policies of the University of Hawaii with those of the University of California, Santa Cruz. 2014 WL 1102660, *6. However, as the Court also pointed out, both GAFE and non-GAFE users may have also “learned of Google’s interceptions” from a “panoply of [other] sources.” Id. at *17. Google’s practices were, for instance, subject to media scrutiny. Id. at n. 12. Thus, “[a] fact-finder, in determining whether [an individual] consented, would have to evaluate to which of the various sources each individual user had been exposed and whether each individual knew about and consented to the interception based on the sources to which she was exposed.” Id. at *18 (internal quotation marks omitted). Sixth, and finally, Plaintiffs’ counsel has had ample time to prepare and look into the specific factual allegations that pertain to each Plaintiff. Over the past five months, Plaintiffs’ counsel has been well aware that the Court may sever Plaintiffs, and that the FAC lacked sufficient factual specificity as to each Plaintiff. Nonetheless, the Court recognizes the unique nature of this case and is cognizant of the inherent difficulties in filing 876 individual complaints. Accordingly, although the Court’s general practice is to provide parties 30 days to file an amended pleading, the Court will grant Plaintiffs’ request for extra time, which Plaintiffs’ counsel made at the August 18, 2016 case management conference. Thus, if Plaintiffs elect to proceed in this litigation, Plaintiffs must file 876 complaints within 45 days of this Order. IV. CONCLUSION For the foregoing reasons, Google’s motions to sever are GRANTED. All but the original three Plaintiffs — Ryan Corley, William Dormann, and Shannon Mehaffey — are SEVERED. Should the other 876 Plaintiffs elect to proceed with their claims, Plaintiffs’ counsel shall file 876 individual complaints. Plaintiff shall do so within 45 of this Order. Failure to meet this 45-day deadline will result in a dismissal of the 876 Plaintiffs’ claims with prejudice. Plaintiffs may not add new causes of actions or parties without leave of the Court or stipulation of the parties pursuant to Federal Rule of Civil Procedure 15. IT IS SO ORDERED. Footnotes [1] All docket citations will be to the Corley case unless otherwise noted. [2] In this Order, the Court abbreviates this requirement as the "same transaction or occur [283] rence” requirement. [3] , The Court notes that the FAC names 200,000 Doe Defendants. The FAC makes no specific allegations as to these Doe Defendants. It simply states that these Defendants are "principals, agents, partners, affiliates, officers, directors, shareholders, creditors, members, employees, managers, joint venturers, co-venturers, and/or co-conspirators ... and were actfing] within the ... scope of their agency” in assisting Google. FAC ¶ 5. The holding in On the Cheap — that 5,000 Doe defendants were unmanageable and were misjoined — may therefore apply with equal or greater force here, where there are 200,000 Doe Defendants. Moreover, under Ninth Circuit law, "[a]s a general rule, the use of 'John Doe’ to identify a defendant is not favored.” Gillespie v. Civiletti, 629 F.2d 637, 642 (9th Cir.1980). Google, however, does not move for misjoinder of Defendants, but reserves the right to do so at a later time. Mot. at 7 n.5. Under these circumstances, the Court does not address misjoinder of Doe Defendants at this time. [4] In In re Google, Inc., 2013 WL 5423918, *14 (N.D.Cal. Sept. 26, 2013), the Court held that "a reasonable Gmail user [who] read” the amendments to the TOS and Privacy Policy on March 1, 2012 “would not have necessarily [have] understood that her emails were being intercepted to create user profiles or to provide targeted advertisements.” [5] The Court notes that Google's briefing does not develop any specific arguments in response to the common question of law or fact requirement. See Reply at 3 (challenging only same transaction or occurrence requirement). Moreover, Google did not challenge commonality in opposing class certification, and some courts have held that Rule 23(a)(2)’s commonality requirement and Rule 20(a)’s common question of law or fact requirement are one and the same. See Lowery v. Ala. Power Co., 483 F.3d 1184, 1202 n. 46 (11th Cir.2007) (treating 23(a)(2) and Rule 20(a) as functionally equivalent) (citations omitted). Under these circumstances, the Court would likely find that Plaintiffs had satisfied Rule 20(a)'s common question of law or fact requirement.