Opinion
22-CV-03580-WHO (VKD)
04-10-2024
IN RE META PIXEL HEALTHCARE LITIGATION This Document Relates To: All Actions.
REDACTED ORDER RE DISCOVERY DISPUTES RE DATA SOURCES AND SAMPLING RE: DKT. NOS. 430, 434, 438
VIRGINIA K. DEMARCHI UNITED STATES MAGISTRATE JUDGE
Plaintiffs and defendant Meta Platforms, Inc. (“Meta”) ask the Court to resolve their disputes concerning plaintiffs' requests for production of class member data and identification of the relevant data sources. Dkt. Nos. 430, 434, 438. The Court held a hearing on these matters on April 9, 2024. Dkt. No. 449. For the reasons stated on the record during the hearing and as further explained below, the Court orders as follows:
11. Class member data. Referencing their Requests for Production Nos. 9-11, 58-61, and 99-101, plaintiffs contend that they have a right to all responsive data for all class members as a matter of law, regardless of the burden of such production, and that any sampling that Meta might propose to address the burden of production would prejudice plaintiffs' ability to move for class certification. Dkt. No. 430 at 1-2. As the Court observed during the hearing, none of the authority on which plaintiffs rely supports this position. Meta argues that plaintiffs' demand for all event data concerning all putative class members would impose an extreme burden on Meta, and it has explained why this is so. Id. at 5. Meta says that it is prepared to discuss a sampling approach with plaintiffs, but plaintiffs have so far refused to entertain any such proposal. Id. at 6-7.
With respect to the (Redacted) data sources the parties agree are relevant, non-duplicative data sources, Meta must identify the data sources from which it would be unduly burdensome to produce all responsive class member data and explain why that is so. For those same data sources, Meta must provide its sampling proposal to plaintiffs by April 17, 2024. The parties must then confer about the proposal, and about any counterproposals or suggestions for mitigating burden made by plaintiffs. The parties must complete their discussions by April 26, 2024. If a dispute remains regarding the production of class member data, the parties shall submit a joint status report to the Court no later than April 30, 2024. The status report shall identify the issues in dispute, without argument, and shall propose an efficient means for presenting the remaining issues to the Court for resolution.
During the hearing, Meta advised that it believes the following Hive tables contain relevant, non duplicative data:
2. Relevant data sources. The parties have yet to agree on the relevant data sources from which Meta will produce responsive data. Meta describes its efforts to identify the relevant data sources and to obtain plaintiffs' agreement regarding the relevant data sources for production, only to be met with plaintiffs' demands for all data from all sources. See Dkt. No. 434 at 1-3. Plaintiffs respond that Meta has repeatedly frustrated their efforts to ascertain which data sources contain relevant, non-duplicative data, and that Meta is responsible for the delay in production of information plaintiffs contend is critical to their case. See Dkt. No. 438 at 1-3. While plaintiffs point to potential data sources from which they contend Meta should produce documents (in addition to those Meta has identified), id. at 3, their complaints about Meta's failure to provide the information they require are not supported by anything in the discovery dispute submission they filed. See generally Dkt. No. 438. Meanwhile, the deadline for substantial completion of document production is April 19, 2024. Dkt. No. 405.
In its November 13, 2023 order denying plaintiffs' motion for discovery sanctions, the Court concluded that Meta's production with respect to Requests for Production Nos. 6 and 7 was incomplete, and ordered as follows:
RFP 6 asks for “[d]ocuments sufficient to identify the databases, logs, or other electronic sources where Meta receives, re-directs, or stores event-level and/or derived data collected through and associated with the Meta Pixel for Medical Provider Web-Properties.” Dkt. No. 380 at 3. RFP 7 asks for “[d]ocuments sufficient to identify and describe each of the fields in databases, logs, or other electronic sources where Meta receives, re-directs, or stores event-level and/or derived data collected through and associated with the Meta Pixel for Medical Provider Web-Properties . . . [including] data dictionaries.” Id.
[T]he Court concludes that the following data sources are responsive to Rfp6 (Redacted)
To the extent these data sources have fields, Meta shall produce documents sufficient to identify and describe each of the fields in response to RFP 7. If Meta contends that its source code is the only place that identifies and/or describes a particular field, Meta shall so state in writing in response to RFP 7.
Plaintiffs have not shown that the following data sources contain Pixel data or are otherwise responsive to RFP 6 or RFP 7, or that Meta failed to produce documents from these sources in violation of the Court's prior order: Everstore, Manifold, TAO, MySQL, ZippyDB/Akkio, Memcache, Laser, Puma, Swift, Stylus, (Redacted), EventScribe, LogDevice, Meta's Data Lakehouse, and “log sources” other than the Hive tables.Dkt. No. 380 at 4. The Court set a deadline of December 11, 2023 for Meta's production of documents responsive to RFPs 6 and 7. Id. During the hearing on the present dispute, plaintiffs' counsel referred repeatedly to this order and argued that Meta had failed to comply with it. However, plaintiffs have not yet raised any issues of non-compliance with the Court, nor did they include a discussion of any such issues in their written discovery dispute submission regarding the present dispute.
As the parties have asked the Court to resolve a dispute about which they have not adequately conferred, and as to which plaintiffs have not even adequately stated their position, the Court will not decide the dispute on the merits as this time. By April 24, 2024, plaintiffs must specifically advise Meta in writing any additional data sources, beyond the (Redacted) tables identified in footnote 1, from which they believe Meta should produce responsive data. If, as plaintiffs argued during the hearing, there is a more efficient and less burdensome means to identify potentially relevant data sources and data fields than by review of Meta's source code, the Court expects Meta to diligently investigate these alternate means. The parties must confer regarding which data sources are relevant and non-duplicative, and if a dispute remains, the parties shall submit a joint status report to the Court no later than May 8, 2024. The status report shall identify the issues in dispute, without argument, and shall propose an efficient means for presenting the remaining issues to the Court for resolution.
E.g., during the hearing, plaintiffs' counsel referred to “data swarm” tables and a “script” for identifying “upstream” and “downstream” tables in Hive, the table fields, and the field descriptions-none of which was addressed in plaintiffs' written submission.
IT IS SO ORDERED.