From Casetext: Smarter Legal Research

Rodriguez v. Google LLC

United States District Court, Northern District of California
Dec 1, 2021
20-cv-04688-RS (AGT) (N.D. Cal. Dec. 1, 2021)

Opinion

20-cv-04688-RS (AGT)

12-01-2021

ANIBAL RODRIGUEZ, et al., Plaintiffs, v. GOOGLE LLC, Defendant.


DISCOVERY ORDER RE: ESI PRESERVATION DISPUTE

Re: Dkt. Nos. 166, 167

ALEX G. TSE, United States Magistrate Judge

Numerous mobile apps use Google's Firebase software; and when people use these apps, their usage data is transmitted to Google. The plaintiffs in this proposed class action used Firebase-powered apps but also turned off a particular Google data-collection feature on their phones or on Google's website. Because they turned this feature off, they assert that Google didn't have consent to collect their app usage data, but did so anyway.

Currently at issue is a dispute that centers on the fact that Google, every 56 days, is deleting the usage data it collects from Firebase-powered apps. The plaintiffs assert that Google is destroying relevant ESI, and they seek a court order that would require Google to stop regularly deleting this data, at least until the parties can agree on a preservation plan. Google, meanwhile, insists that its retention policy is reasonable because the costs of preserving more than 56 days' worth of this data greatly exceed the benefits.

Having considered the evidence, the Court agrees with Google. Google collects an immense amount of data from Firebase-powered apps. Each day, the company's app measurement logs record, on average, 135 billion entries from Firebase-powered apps, with those entries occupying over 27 petabytes of storage space (one petabyte is equivalent to one million gigabytes). See Dkt. 166-3, Ganem Decl. ¶¶ 3-4. “Using public storage cost figures, ” Google estimates that “storing two years' worth of such data would cost $3 million, and three years would cost $6 million, ” sizeable sums that would be above and beyond ordinary litigation costs. Id. ¶ 5.

As the plaintiffs note, these estimates are based on what it would cost Google to store all Firebase app usage data during the given periods, not just the Firebase app usage data that is tied to the proposed class members (i.e., app users who turned off a particular Google data-collection feature). But if Google were required to store the latter, smaller category of app usage data, the company asserts that it would also need to store the former, larger category, because “[m]any users turn [the data-collection feature] on and off repeatedly, and the logs are anonymized, so only preserving them in their entirety would enable Google to query them for particular . . . users in particular moments in time.” Dkt. 167 at 6. No reason for doubting this plausible explanation has been given, so the Court accepts Google's estimates as true.The motion to seal the estimates and declaration offered in support of them (Dkt. 166) is denied. Google hasn't convinced the Court that “specific prejudice or harm will result” if this information is publicly disclosed. Phillips ex rel. Ests. of Byrd v. Gen. Motors Corp., 307 F.3d 1206, 1210-11 (9th Cir. 2002). Google must file unreacted public versions of the documents at issue by next Wednesday.

On the other end of the scale, the plaintiffs haven't persuasively explained why they need more than 56 days' worth of the data in question. They explain that they seek the data “in connection with their class certification motion and for the parties' experts, to quantify the number of illegal interceptions and Google's unjust enrichment.” Id. at 3. But they don't suggest that they intend to, or have the means to, sift through and count 27 petabytes of data per day. Instead some form of statistical sampling will likely be necessary, and the plaintiffs haven't explained why they can't conduct such a sampling using the data Google regularly saves, or using the named plaintiffs' data (which Google has offered to save and produce) as opposed to all of the proposed class members' data (which is what the plaintiffs seek).

To pin down the number of times Google has intercepted the plaintiffs' data, and the amount by which Google has profited from that data, the plaintiffs may also depose Google and its employees and propound written discovery requests. These methods, plus some form of statistical sampling, would be, as Google correctly puts it, “more direct and far less burdensome than filling a warehouse with massive server cabinets full of anonymous individuals' data.” Id. at 7.

Rule 37, as Google notes, “does not call for perfection” in preserving ESI; it calls for “reasonableness.” Adv. Comm. Notes to 2015 Amend. of Fed.R.Civ.P. 37(e). And whether preservation is reasonable “depends on whether what was done-or not done-was proportional to [the] case.” Pettit v. Smith, 45 F.Supp.3d 1099, 1107 (D. Ariz. 2014) (emphasis omitted) (quoting Rimkus Consulting Grp. v. Cammarata, 688 F.Supp.2d 598, 613 (S.D. Tex. 2010)). Here, the plaintiffs haven't persuaded the Court that Google's current retention policy for the app usage data in question is unreasonable. This isn't to say that the data is irrelevant, or that Google doesn't need to retain and turn over a subset of it. But the Court won't require Google to start saving petabytes of data per day, indefinitely, without a more compelling showing of need.

* * *

Related to the ESI dispute just addressed, the plaintiffs make two other requests. First, they seek a court order that would require Google to identify “any relevant data sources” related to Google's collection and use of the proposed class members' app usage data. Dkt. 167 at 4. Second, they seek a court order that would require Google to negotiate an ESI preservation plan. On the current record, neither request will be granted.

As to the first, court intervention would be premature. The plaintiffs must first utilize all discovery tools at their disposal, including Rule 30(b)(6) depositions, to attempt to identify the data sources they seek. It doesn't appear that they have done so to date, and until they do the Court won't intervene. As to the second request, the parties have already stipulated to certain ground rules for ESI discovery, see Dkt. 73, and it seems that what the plaintiffs want now is an ESI preservation plan that would require Google to preserve more than 56 days' worth of app usage data for all putative class members. The Court won't require Google to negotiate such a preservation plan for the reasons stated above.

The parties must continue to cooperate in good faith to preserve and share relevant ESI. But the Court won't require Google to modify its current ESI retention policy.

The plaintiffs' requests for relief are denied.

IT IS SO ORDERED.


Summaries of

Rodriguez v. Google LLC

United States District Court, Northern District of California
Dec 1, 2021
20-cv-04688-RS (AGT) (N.D. Cal. Dec. 1, 2021)
Case details for

Rodriguez v. Google LLC

Case Details

Full title:ANIBAL RODRIGUEZ, et al., Plaintiffs, v. GOOGLE LLC, Defendant.

Court:United States District Court, Northern District of California

Date published: Dec 1, 2021

Citations

20-cv-04688-RS (AGT) (N.D. Cal. Dec. 1, 2021)