Todd Crowder, et al. v. LinkedIn Corporation

Opinion

1 2 3 4 5 6 7 8 UNITED STATES DISTRICT COURT 9 NORTHERN DISTRICT OF CALIFORNIA 10 San Francisco Division 11 TODD CROWDER, et al., Case No. 22-cv-00237-HSG (LB)

12 Plaintiffs, DISCOVERY ORDER 13 v. Re: ECF Nos. 162, 164, 167 14 LINKEDIN CORPORATION, 15 Defendant. 16 17 INTRODUCTION 18 The parties have discovery disputes in this putative class action, where the plaintiffs claim 19 monopolization and attempted monopolization from practices that allegedly allow LinkedIn to 20 overcharge Premium subscribers, in violation of § 2 of the Sherman Act. The two categories of 21 alleged anticompetitive conduct are (1) offering potential rivals access to LinkedIn’s private user 22 data through application-programming interfaces (API) unavailable to others, on the condition that 23 the rivals don’t compete with LinkedIn, and (2) integrating LinkedIn’s user data with parent 24 company Microsoft’s Azure cloud product, thereby “tying up and driving up prices for scarce 25 hardware resources.”1 26

27 1 Order – ECF No. 148 at 1; Disc. Letter Brs. – ECF Nos. 162, 164, 167. Citations refer to the 1 Three discovery letter briefs are before the court. The first (ECF No. 162) concerns the 2 custodians for LinkedIn’s production in response to the plaintiffs’ requests for production. The 3 second (ECF No. 164) concerns the scope of production for LinkedIn’s RFPs requesting that the 4 plaintiffs search their personal social-media accounts and produce professional-networking 5 communications and contact lists. The third (ECF No. 167) concerns the sufficiency of LinkedIn’s 6 document production of sales, pricing, and financial data.2 7 For ECF No. 162, LinkedIn must produce documents from the following disputed custodians in 8 addition to the sixteen custodians LinkedIn has already identified: Ryan Roslansky, Daniel Shapero, 9 and Erran Berger. For now, the plaintiffs’ request to add additional custodians — Joseph Baribeau, 10 Ines Gonzalez, Allison Guillen-Capo, Hahn Vo, Blake Lawit, Vivek Hariharan, and Dhwani Shah 11 — is denied without prejudice (except the conceded discovery for Guillen-Capo). For RFP Nos. 13– 12 16, LinkedIn must produce documents sufficient to show the scope and status of the Azure 13 migration. The parties must meet and confer regarding RFP Nos. 25–26 after LinkedIn serves its 14 responses and objections. 15 For ECF No. 164, the court grants LinkedIn’s request in part. The plaintiffs must conduct a 16 limited, targeted search of a subset of their social-media accounts, as described below. The court 17 denies LinkedIn’s request in RFP No. 13 for production of the plaintiffs’ complete connections 18 and follower lists. 19 For ECF No. 167, LinkedIn has already agreed to produce what exists back to the date that the 20 plaintiffs seek, January 1, 2014. This is sufficient. 21 22 ANALYSIS 23 Parties may obtain discovery regarding any nonprivileged matter that is relevant to any party’s 24 claim or defense and proportional to the needs of the case, considering the importance of the 25 issues at stake in the action, the amount in controversy, the parties’ relative access to relevant 26 information, the parties’ resources, the importance of the discovery in resolving the issues, and 27 1 whether the burden or expense of the proposed discovery outweighs its likely benefit. Information 2 within this scope of discovery need not be admissible in evidence to be discoverable. Fed. R. Civ. 3 P. 26(b). “Pretrial discovery is ordinarily accorded a broad and liberal treatment.” Peng v. Nw. 4 Mut. Life Ins. Co., No. 17-cv-01760-SI, 2017 WL 3007030, at *1 (N.D. Cal. July 14, 2017) 5 (cleaned up) (quoting Shoen v. Shoen, 5 F.3d 1289, 1292 (9th Cir. 1993)). Moreover, “[t]he test 6 for relevance is not overly exacting: evidence is relevant if it has ‘any tendency to make . . . more 7 or less probable . . . [a] fact [that] is of consequence in determining the action.’” In re Volkswagen 8 “Clean Diesel” Mktg., Sales Pracs., & Prods. Liab. Litig., MDL No. 2672 CRB (JSC), 2018 WL 9 340640, at *1 (N.D. Cal. Jan. 9, 2018) (quoting Fed. R. Evid. 401). 10 The party moving to compel discovery “has the initial burden of establishing that the 11 information sought is relevant to any party’s claim or defense and proportional to the needs of the 12 case.” Impinj, Inc. v. NXP USA, Inc., No. 19-cv-03161-YGR (AGT), 2022 WL 16586886, at *2 13 (N.D. Cal. Nov. 1, 2022) (cleaned up). The party resisting discovery bears the burden of showing 14 that the discovery should not be allowed and of supporting its objections with competent evidence. 15 Lofton v. Verizon Wireless (VAW) LLC, 308 F.R.D. 276, 281 (N.D. Cal. 2015). Under Rule 26(b), 16 the court must limit discovery that is “unreasonably cumulative or duplicative,” obtainable from a 17 less burdensome source, or where the burden “outweighs its likely benefit.” Fed. R. Civ. P. 18 26(b)(1), (b)(2)(C). 19 The responding party generally is “in a better position” to identify those employees “most 20 likely to possess responsive ESI.” Emerson v. Iron Mountain Info. Mgmt. Servs., Inc., No. 20-cv- 21 08607-YGR (AGT), 2021 WL 8085488, at *1 (N.D. Cal. Sep. 2, 2021). That said, a requesting 22 party may obtain additional custodians upon a showing that the disputed custodians “possess 23 uniquely relevant information that is not available from the sources already designated.” In re 24 Facebook, Inc. Consumer Priv. User Profile Litig., No. 3:18-MD-02843-VC-JSC, 2021 WL 25 10282213, at *1 (N.D. Cal. Nov. 14, 2021). 26 27 1 1. Custodians for LinkedIn’s Document Production (ECF No. 162) 2 The RFPs at issue are at ECF No. 163-1. The plaintiffs group them into four categories: 3 (1) Market definition (or relevant market). Information related to (1a) LinkedIn’s actual or potential competition in the Professional Social 4 Networking (PSN) Market and (1b) LinkedIn’s PSN product and how it is offered and sold. (RFP Nos. 1, 5, 10, 18, 24–26.) 5 (2) Monopoly power. Information related to LinkedIn’s market share, the barriers 6 to entry surrounding the PSN Market, (2a) the Data, Machine Learning, and Inference Barrier to Entry (DMIBE), and (2b) the impacts of LinkedIn’s 7 conduct on its competitive position (anticompetitive impact). (RFP Nos. 6(b), 8, 8 13, 24–27.) (3) Anticompetitive conduct. Information related to (3a) LinkedIn’s specific intent 9 to monopolize the PSN Market (including LinkedIn’s AI/machine learning and 10 infrastructure strategy and efforts to strengthen the DMBIE), (3b) LinkedIn’s private API program and its creation and/or maintenance of monopoly power in 11 the PSN Market (including whether LinkedIn intended to exclude rivals and forestall/decrease competition through the program), and (3c) LinkedIn’s 12 agreement with Microsoft related to Azure integration (including whether below-cost pricing for Azure compete impacted PSN competition and thereby 13 contributed to LinkedIn’s monopolization/attempted monopolization. alleged 14 anticompetitive conduct. (RFP Nos. 1–4, 6(a), 7–8, 12–17, 19, 25–27.) (4) Remedy/ damages. The RFPs concern the overcharge injury to LinkedIn 15 Premium subscribers from the monopolization/attempted monopolization and 16 seek (4a) financial details of LinkedIn’s business and (4b) information related to LinkedIn Premium Pricing. (RFP Nos. 9–11, 22–23.)3 17 LinkedIn has agreed to produce documents from sixteen custodians.4 The plaintiffs seek to add 18 ten more: Joseph Baribeau, Erran Berger, Ines Gonzalez, Allison Guillen-Capo, Vivek Hariharan, 19 Blake Lawit, Ryan Roslansky, Dhwani Shah, Daniel Shapero, and Hahn Vo.5 LinkedIn must 20 produce documents from the following disputed custodians: Ryan Roslansky, Daniel Shapero, and 21 Erran Berger. For now, the plaintiffs’ request to add additional custodians — Joseph Baribeau, 22 Ines Gonzalez, Allison Guillen-Capo, Hahn Vo, Blake Lawit, Vivek Hariharan, and Dhwani Shah 23 — is denied without prejudice (except the conceded discovery for Guillen-Capo). 24 The next sections address the disputed custodians in groups organized by the relevant issues. 25 26 3 Disc. Letter Br. – ECF No. 162 at 2. 27 4 Id. at 4. 1 1.1 Ryan Roslansky (CEO/former CPO/Microsoft EVP) (topics 1, 1a, 1b, 2, 2a, 2b, 3a, 3b, 2 3c, 4a, and 4b) & Daniel Shapero (COO/interim CPO) (topics 1, 1a, 1b, 2, 3, 3a, 3b, and 3c) 3 Summary of dispute. The plaintiffs contend that Roslansky’s files are uniquely relevant because 4 he was CPO for the launch and development of the private API program and the decision to 5 integrate with Azure. He now has a “dual-hat” role as LinkedIn CEO and Microsoft EVP. Shapero 6 is COO and interim CPO, leading sales, operations, marketing, pricing, and strategic planning, and 7 served as VP of Product (2014–2019) and CBO (2019–2021).6 LinkedIn responds that neither 8 executive was named in the complaint, their files will contain excessive irrelevant information, and 9 the ten already-designated business leaders will have the relevant competition documents.7 10 Analysis. In an antitrust monopolization case, the CEO and COO are not peripheral figures. 11 Roslansky oversaw both pillars of the alleged anticompetitive conduct — the API program (as 12 CPO from 2009–2020) and the Azure integration decision — and his current dual role at LinkedIn 13 and Microsoft makes his files uniquely likely to contain documents about the competitive 14 relationship between the two companies that lower-level custodians would not possess. Shapero 15 similarly occupied the CPO, CBO, and COO chairs across the relevant period, giving him a span 16 of oversight across pricing, competitive strategy, and business development that differs in kind 17 from any single business-unit leader. 18 LinkedIn’s objection that these executives’ files will contain “excessive irrelevant information” 19 is a proportionality concern that can be addressed through appropriate search terms, date 20 restrictions, and the parties’ existing ESI protocol — not by excluding the custodians altogether. 21 The fact that neither executive was named in the complaint is immaterial; the question is whether 22 their files are likely to contain responsive, non-duplicative documents, and the answer is yes. Both 23 individuals occupied positions that gave them a unique, enterprise-wide view of competitive 24 strategy, API policy, Azure integration, and Premium pricing that the designated business-unit 25 leaders, by definition, do not share. 26

27 6 Id. at 3. 1 LinkedIn must add Roslansky and Shapero as custodians. 2 1.2 Erran Berger (topics 1, 2, 3, 3a, and 4) 3 Summary of dispute. The plaintiffs identify Berger as the leader of “engineering for 4 LinkedIn’s complete product portfolio,” including Premium, who led “all growth, engagement and 5 monetization aspects of LinkedIn’s consumer experience,” including “on the monetization side . . . 6 [LinkedIn’s] Premium subscription.”8 LinkedIn argues that Berger’s “highly technical files” are 7 irrelevant and disproportionate, particularly in connection with the DMIBE-related RFP Nos. 25– 8 26, and flags that its responses to those RFPs are not yet due.9 9 Analysis. Berger’s role is not limited to the technical subjects LinkedIn emphasizes. The 10 plaintiffs assert, and LinkedIn does not counter, that Berger led engineering for the entire product 11 portfolio, including Premium, and oversaw growth, engagement, and monetization. These 12 responsibilities are directly relevant to the monopoly-power and anticompetitive-conduct issues 13 (topics 1–4). His perspective bridges the technical and business dimensions of LinkedIn’s product 14 strategy in a way that the designated business-development or Premium-product custodians do not. 15 The court orders LinkedIn to add Berger as a custodian for the RFPs at issue in this letter brief. 16 The request is premature for RFP Nos. 25–26 (DMIBE) because LinkedIn’s responses and 17 objections to those RFPs are not yet due. 18 1.3 Ines Gonzalez, Allison Guillen-Capo, Blake Lawit, and Hahn Vo 19 Summary of dispute. The plaintiffs seek to add LinkedIn’s General Counsel (Lawit) and three 20 additional attorneys (Gonzalez, Guillen-Capo, and Vo) as custodians. The plaintiffs argue these 21 attorneys are relevant because they are sole signatories to certain private API agreements, and 22 their files may contain documents bearing on LinkedIn’s intent in structuring and enforcing those 23 agreements.10 LinkedIn responds that while the proposed attorney-custodians signed the 24 agreements, they did not negotiate them and are merely employees with authority to sign. It is 25 26 8 Id. at 2 (cleaned up). 27 9 Id. at 5. 1 already offered to produce external communications between the LinkedIn attorneys who 2 negotiated the API agreements and the API partners (Allison Guillen-Capo,11 Jane Slater, and 3 Christopher Grant). They, working with the business-development team, are responsible for 4 negotiating all private API agreements and have been doing so since 2014, 2015, and 2017, 5 respectively. The documents would yield little non-privileged, non-duplicative material.12 6 Analysis. The court denies this request without prejudice. LinkedIn has already offered to 7 produce external communications between its API-negotiating attorneys and API partners. The 8 requesting party bears the burden of showing that additional attorney custodians possess “uniquely 9 relevant information that is not available from the sources already designated.” In re Facebook, 10 2021 WL 10282213, at *1. The plaintiffs have not met that burden here. That these attorneys 11 signed the API agreements does not, without more, establish that their custodial files will contain 12 non-privileged, non-duplicative responsive documents beyond those produced by the business- 13 development custodians and the existing attorney-custodians. 14 This ruling is without prejudice. If, after reviewing LinkedIn’s production, the plaintiffs 15 identify specific gaps that only these attorney custodians can fill, they may renew the request with 16 a particularized showing. 17 1.4 Vivek Hariharan (topics 2, 2a, 2b, 3, 3a, 3b, and 3c) 18 Summary of dispute. The plaintiffs describe Hariharan as Engineering Director, AI Inference, 19 Search & Recommendation, and was the Engineering Director, Feed Infrastructure. He “leads 20 LinkedIn’s AI Platform organization responsible for all online inference, retrieval, and LLM/SLM 21 serving across Feed, Ads, Jobs, and People Search” and “established the architectural foundation 22 for LinkedIn’s current AI platform.” The plaintiffs argue that he is relevant to monopoly power, 23 barriers to entry (including the DMIBE), and anticompetitive conduct.13 LinkedIn contends his 24 “highly technical files” are irrelevant and disproportionate.14 25 26 11 This at least in part falls within the plaintiffs’ RFP. 12 Id. at 4. 27 13 Id. at 2–3. 1 Analysis. The court denies this request without prejudice. Hariharan’s role is predominantly 2 technical (AI inference, retrieval, and LLM/SLM serving infrastructure). The discovery is about 3 the terms of API agreements and the competitive effects of Azure integration and pricing, not the 4 engineering architecture of LinkedIn’s AI systems. The plaintiffs’ DMIBE theory may ultimately 5 require technical discovery, but the request is premature because the responses are not yet due. At 6 this stage, the plaintiffs have not demonstrated that Hariharan’s files will yield responsive, non- 7 duplicative documents for the currently ripe RFPs. 8 1.5 Dhwani Shah (topics 2, 2b, 3, 3a, and 3b) 9 Summary of dispute. The plaintiffs identify Shah as Director, Partner Engineering, leading a 10 team focused on external developers and partner-solution engineers, who “provided product 11 leadership for LinkedIn’s members first data platform.”15 LinkedIn responds that the “technical 12 implementation” of the API program is not at issue because the plaintiffs challenge the contractual 13 terms of the agreements, not their technical execution.16 14 Analysis. The court denies this request without prejudice. The plaintiffs’ claims target the API 15 agreement terms, not the technical infrastructure through which access was provided. The 16 designated business-development custodians and external attorney communications are more 17 directly relevant to how API agreements were negotiated, structured, and enforced. If the 18 plaintiffs’ review of the production reveals that technical-implementation documents are 19 necessary, they may renew this request. 20 1.6 Joseph Baribeau (topics 3, 3a, 3c, 4, and 4a) 21 Summary of dispute. The plaintiffs assert that Baribeau was the “finance” representative of the 22 “core team” for the Blueshift aspect of the Azure migration and should be a custodian for topics 23 related to anticompetitive conduct and financial information, including the Azure migration.17 24 LinkedIn argues that the Azure migration was never completed — it was paused — and that a 25 26 15 Id. at 3 (cleaned up). 27 16 Id. at 4. 1 custodial search for the finances of a migration that “never occurred” is disproportionate. LinkedIn 2 is producing documents to show its use of Azure and that the migration was never restarted.18 3 Analysis. The court denies the plaintiffs’ request to add Baribeau as a custodian. Whether the 4 Azure migration was completed, paused, or partially implemented is itself a factual dispute, but 5 LinkedIn’s offer to produce documents sufficient to show the extent of its migration to Azure and 6 to confirm the migration’s status is an adequate starting point, as discussed in the next section. If 7 LinkedIn’s production proves inadequate, the plaintiffs may seek further relief, including the 8 addition of Baribeau or another custodian with knowledge of Azure-related finances. 9 1.7 Azure Migration (RFP Nos. 13–16) 10 The First Amended Complaint (FAC) alleges that LinkedIn anticompetitively integrates with 11 Microsoft Azure.19 Whether that migration was completed, paused, or partially executed is a 12 disputed factual question relevant to the plaintiffs’ § 2 claims. LinkedIn submitted an email 13 supporting its point.20 The plaintiffs are entitled to test what actually happened through discovery. 14 See, e.g., Strike 3 Holdings, LLC v. John Doe Subscriber, No.: 21-cv-63-AJB-AHG, 2022 WL 15 22898319, at *3 (S.D. Cal. July 23, 2022). At the same time, designating custodians — at least 16 before preliminary document discovery — could be disproportionate if the migration was halted. 17 LinkedIn will produce documents showing (1) the migration was never restarted and (2) the 18 extent of its use of Azure. It suggests a Rule 30(b)(6) deposition.21 For now, the court orders that 19 compromise, with the proviso that there should be a full illumination through the document 20 production of the core requests in RFP Nos. 13–16 and an illustration of LinkedIn’s representation 21 that there is no point in designating a custodian for the finances of a migration that “never 22 occurred.”22 Then, a Rule 30(b)(6) deposition on the Azure topic can illuminate the document 23 production. Any further discovery issues can be taken up after that deposition. 24

25 18 Id. at 5. 26 19 FAC – ECF No. 65 at 57–65 (pp. 54–62) (¶¶ 279–325). 20 Email, Ex. B to Disc. Letter Br. – ECF No. 161-4. 27 21 Disc. Letter Br. – ECF No. 162 at 5. 1 1.8 DMIBE-Related Requests (RFP Nos. 25–26) 2 LinkedIn’s responses and objections to RFP Nos. 25–26 are not due until April 6, 2026.23 Any 3 dispute is premature. 4 5 2. LinkedIn’s RFPs for the Plaintiffs’ Social-Media Discovery (ECF No. 164) 6 LinkedIn seeks two categories of discovery from the plaintiffs: (1) for RFP No. 12, 7 communications on social-networking platforms related to professional topics; and (2) for RFP 8 No. 13, documents sufficient to identify the plaintiffs’ connections and followers on those 9 platforms: 10 RFP No. 12: All communications sent or received by You on any Social Networking Platform to 11 the extent such communications are related to employment, career, occupation, hiring, selling, marketing, professional development, recruitment, professional 12 networking, marketing, branding, professional training, or any other profession- 13 related topics. RFP No. 13: 14 Documents sufficient to identify each of the individuals, entities, corporations, or 15 other accounts with which You are connected, followed by, follow, or otherwise linked on any Social Networking Platform.24 16 It asserts that “crucial” to the plaintiffs’ monopoly claim is the allegation that other social 17 networks (Facebook, Twitter/X, YouTube, and the like) do not compete with LinkedIn and do not 18 provide like business or professional connections.25 19 LinkedIn limited its request to RFP No. 12 to “private, direct, and public postings, messages, 20 and chats” (including postings where the plaintiffs are tagged, referenced, or appeared, but not 21 including the plaintiffs’ “likes” or similar reactions to posts).26 For RFP No. 13, it wants 22 “documents sufficient to show Plaintiffs’ connections on social-media platforms.”27 23 24

25 23 Id. 26 24 RFPs, Ex. A to Disc. Letter Br. – ECF No. 164-1 at 2–3. 25 Disc. Letter Br. – ECF No. 164 at 1 (cleaned up). 27 26 Id. 1 2.1 RFP No. 12 2 Summary of dispute. LinkedIn contends that substitute products must be considered alongside 3 the allegedly monopolizing product. The plaintiffs’ use of other platforms for professional 4 networking is central to market definition, and interrogatory responses are no substitute for 5 document production. LinkedIn has proposed search terms and asserts that the burden on the 6 plaintiffs is minimal.28 The plaintiffs respond that (1) their verified interrogatory responses 7 uniformly confirm they do not use other platforms for professional purposes, (2) LinkedIn has 8 identified no contrary evidence in the plaintiffs’ 100,216-document production (comprising 9 430,681 pages), (3) LinkedIn’s 127 proposed search terms are “functionally unbounded and overly 10 burdensome,” and (4) the technical burden of searching platforms with limited export functionality 11 is substantial because the platforms generally have no mechanism to export private messages, 12 communications history, or posts from third parties tagging the plaintiffs.29 13 Governing law. Market definition is a central issue in a § 2 case, and “substitute products must 14 also be considered.” United States v. Grinnell Corp., 384 U.S. 563, 571 (1966). Evidence of how 15 consumers actually use competing platforms is relevant to that inquiry. See Fount-Wip, Inc. v. 16 Reddi-Wip, Inc., 568 F.2d 1296, 1301 (9th Cir. 1978) (the relevant product market “depends upon 17 the availability of alternative commodities for buyers”). At the same time, courts “have demanded 18 a pretty solid showing of relevance and proportional scope before requiring a litigant to turn over 19 his or her social media posts.” Brown v. City of Crescent City, No. 18-cv-07826-EMC (TSH), 20 2021 WL 857694, at *1 (N.D. Cal. Mar. 8, 2021). And discovery must be proportional to the 21 needs of the case, considering the burden on the producing party and the importance of the 22 discovery in resolving the issues. Fed. R. Civ. P. 26(b)(1). 23 Analysis. Interrogatory responses are not a substitute for document production. But this is not a 24 case in which the interrogatory responses are evasive or incomplete. The plaintiffs identified every 25 social-networking platform they use and described the purpose of their use. Schulte affirmed that he 26

27 28 Id. at 2. 1 does not use his other platforms “for business or professional purposes.” Vance and Crowder 2 described exclusively personal activities: keeping up with friends and family, viewing photos, 3 pursuing hobbies, and following current events.30 LinkedIn’s observation — that Vance and 4 Crowder did not aver that they do not use their social-media accounts “for business or professional 5 purposes” 31 — is unpersuasive given the verified interrogatory responses and LinkedIn’s failure to 6 identify any documents from the plaintiffs’ email production (based on LinkedIn’s search terms) 7 suggesting use of social media for professional purposes.32 8 That said, LinkedIn, like the plaintiffs, is entitled to test the plaintiffs’ discovery responses 9 through additional discovery. Strike 3, 2022 WL 22898319, at *3. But the 127 search terms 10 include “relationship,” “money,” “sale*,” “skill*,” and “benefits.”33 Those terms encompass 11 plainly personal content with no connection to professional networking. 12 The parties do not make it easy to sort out a proportional discovery response. The court orders 13 the following to test the interrogatory responses. 14 First, any discovery is limited to platforms that plausibly could be used for professional 15 networking. The plaintiffs identified social-media platforms including Facebook, Pinterest, 16 Instagram, TikTok, Discord, Reddit, YouTube, Meetup, Tumblr, Google+, Doximity, and 17 Twitter/X.34 Some platforms seem unlikely to be used for professional networking purposes. 18 Second, LinkedIn’s 127 terms are excessive. 19 Third, given this context, the parties must confer within twenty-one days to address the dispute 20 and must consider starting with a platform that reasonably might reflect the plaintiffs’ actual 21 professional networking activity (hiring, recruitment, job seeking, networking, employment, and 22 the like). Probably this would not include platforms like YouTube or TikTok. If, as the plaintiffs 23 say, there are no export or search functionalities, the parties must confer about whether there are 24

25 30 Id. 26 31 Id. at 2. 32 Id. at 4. 27 33 Id. 1 technical solutions to this problem. LinkedIn has offered to troubleshoot difficulties.35 It is 2 impossible to evaluate burden, despite LinkedIn’s representation that burden is minimal.36 3 2.2 RFP No. 13: Connection and Followers 4 Summary of dispute. LinkedIn requests “documents sufficient to show Plaintiffs’ connections 5 on social-media platforms.”37 The plaintiffs respond that disclosing every Facebook friend, 6 Instagram follower, Reddit connection, Discord server member, and YouTube subscription has no 7 bearing on whether LinkedIn has monopoly power in the PSN Market. They note that LinkedIn 8 has complete data on the plaintiffs’ LinkedIn usage, connections, and activity and “is the entity 9 best positioned to demonstrate the value of its own platform relative to alternatives.”38 10 Analysis. The court denies this request. The marginal relevance of a complete list of the 11 plaintiffs’ social-media contacts across dozens of platforms is low in the context of this case. The 12 plaintiffs’ verified interrogatory responses already describe the purpose of their use of each 13 platform. LinkedIn has identified no basis to believe that connection lists — as distinct from 14 communications — would materially advance the market-definition inquiry that is at issue in the 15 case. The privacy burden of disclosing hundreds (or more) personal contacts on platforms used for 16 friends, family, hobbies, and entertainment substantially outweighs the likely benefit. See Fed. R. 17 Civ. P. 26(b)(1); cf. Kellman v. Spokeo, Inc., No. 3:21-cv-08976-WHO, 2023 WL 411353, at *1 18 (N.D. Cal. Jan. 25, 2023). If the targeted search ordered under RFP No. 12 reveals that any 19 plaintiff does in fact maintain a professional network on another platform, LinkedIn may renew 20 this request as to that platform. 21 22 23 24 25 26 35 Id. 36 Id. 27 37 Id. at 3. ] 3. The Plaintiffs’ RFPs for the LinkedIn’s Financial Data (ECF No. 167) 2 RFP Nos. 9 and 10 seek documents about pricing and prices paid. LinkedIn has pricing data 3 only back to August 2017. It will run pricing terms across the files of several custodians and 4 || produce pricing from January 2014 to August 2017.°? RFP No. 11 seeks documents about costs, 5 || revenues, and profits on U.S. sales of LinkedIn Premium services or any upgraded account 6 || features. LinkedIn will produce Premium and Flagship profit and loss information from January 1, 7 || 2017, to present, which is the only period for the data, and will produce company-wide profit and 8 loss information and Premium-specific revenue for January 1, 2014, to January 1, 2017.*° 9 LinkedIn’s compromise is an acceptable resolution of this dispute. 10 11 CONCLUSION a 12 This resolves ECF Nos. 162, 164, and 167.

2 13 IT IS SO ORDERED.

Dated: April 6, 2026 LAE

LAUREL BEELER A 16 United States Magistrate Judge

Oo Z 18 19 20 21 22 23 24 25 26 27 || 30H: Disc. Letter Br. — ECF No. 167 at 3. 28 || 4° Jd. at 4-5.