In re Google RTB Consumer Privacy Litigation

Background

The plaintiffs and Google asked the court to resolve a discovery dispute about Google’s compliance with an August 26, 2022 order concerning data about the named plaintiffs. That earlier order required Google to produce documents sufficient to show specified categories of information, including sign-up information, consent-related data, settings, activity records, linking-account information, advertisements shown, and cookie-matching information.

The plaintiffs argued that Google’s production was incomplete. They contended that Google had focused on one data source while failing to search other logs, had omitted relevant data fields, had limited production to six weeks, and had produced the data without enough context or structure. They also sought broader information related to information shared with participants in real-time bidding auctions.

Google maintained that it had complied with the August 26 order and had produced named-plaintiff data from multiple time periods and log sources. It argued that many of the additional data fields were irrelevant or reflected information Google maintained internally rather than information shared with real-time bidding participants. Google also said that producing data for the entire class period from sources requiring notice to advertising customers would create an undue burden.

Court’s analysis

The court rejected the plaintiffs’ request for wholesale production of all data fields from the additional logs. It found that the plaintiffs had not shown that producing every field was relevant and proportional to the needs of the case. The court also found unpersuasive the argument that Google’s method of querying relevant fields and outputting the results to files made the production unusable. The court concluded that the plaintiffs had not shown how Google’s method prevented them from understanding or using the data.

The court determined, however, that Google had failed to comply with the August 26 order regarding “verticals.” The opinion describes verticals as interest-based categories associated with account holders. The court said Google had acknowledged sharing at least one related data field with real-time bidding participants during part of the relevant period, although Google said that field was no longer shared beginning around February 2020. Google had not produced documents sufficient to show the detected verticals shared for each named plaintiff or investigated whether that information could be obtained from existing sources.

The court agreed that the required production had to address both explicit sharing of verticals associated with named plaintiffs and information that could allow participants to infer account-holder information based on verticals linked to those plaintiffs.

On time sampling, the court accepted Google’s representation that producing data for the entire class period from sources requiring notice to advertising customers would impose an undue burden. The court held that sampling was appropriate for those sources, but any sample had to be representative. Google had to explain how it selected the six weeks it produced, and the parties had to confer about any disagreement. For sources that did not require notice to advertising customers, the court was not persuaded that full-period production would impose an undue burden.

The court also ruled that requests for broader information “related to” or “in furtherance of” the real-time bidding process went beyond the scope of the August 26 order.

Disposition

The court ordered Google to produce the specified “verticals” information by January 31, 2023. It ordered the parties to confer about the time-sampling issue by January 13, 2023, and allowed them to submit any remaining dispute to the court. The court denied all other relief requested. The opinion was signed by Virginia K. Demarchi, United States Magistrate Judge.