What do you do when the data is changing? – eDiscovery of Dynamic Data

Dynamic data is “information [that] is constantly being updated in real time.” Saturn Sys. v. Militare, 252 P.3d 516, 523 (Col. App. 2011). Dynamic data can include information about badge swipes, log-ins and log-outs, interaction with particular servers or files, mouse clicks, and keystrokes. It may be scattered across servers, including third-party servers. In some cases, it may only be accessible at very high costs, such as the complete suspension of business operations. And even then, it may not be readable or useable without purchasing specific software.

The key for companies and their counsel is to identify the existence of dynamic data early in discovery and raise it affirmatively with opposing counsel as a part of the Rule 26(f) conference. Outside counsel should be prepared to explain what the data source is, the type of data it contains, how costly it would be for the company to collect or preserve that data, and whether the data could be easily produced in discovery. By providing this information, outside counsel can often reach a quick agreement that the data does not need to be preserved, collected, or produced in discovery.

Building up a record regarding the accessibility, usefulness, and cost of dynamic data discovery can also pay dividends in discovery motions practice. For example, in Lickerish, Inc. v. PicClick, Inc., No. 5:22-MC-80152, 2022 WL 2812174 (N.D. Cal. July 18, 2022), PicClick operated a tool that allowed users to search eBay while simultaneously displaying images of celebrities. Id. at *1. Lickerish claimed that PicClick had violated its copyrights over those images. Id. However, the allegedly infringing images were stored on eBay’s servers, not PicClick’s. Id. Lickerish issued a subpoena to eBay seeking information about “the operation of the storage of the allegedly infringing photos on eBay’s servers and caching to PicClick’s network.” Id. at *2.

However, while interactions with “ebay.com” URLs were recorded by eBay, eBay had no way of knowing “what happens when a user clicks [PicClick] links,” short of “click[ing] each of the link[s] and then check[ing] its internal logs to see if there was a communication with eBay’s servers at the same time.” Id. (internal quotation marks omitted). To complicate matters further, eBay used Content Delivery Networks (“CDNs”) to cache images from its servers and display them to users, and so, as a practical matter, eBay would have been required to manually “check each CDN’s log after clicking each PicClick link,” which would have been “extraordinarily burdensome,” as each individual CDN “receive[d] more than 40 billion [communications] each day.” Id. at *1–2. There was also uncertainty about the data’s accuracy, since eBay did not keep logs of every type of communication with the CDNs, so it would have theoretically been possible for communications to be made “without logs reflecting them.” Id. at *2.

Lickerish filed a motion to compel. In its analysis, the court denied the motion, finding that a subpoena seeking discovery of this dynamic data imposed an “undue burden that outweighs the relevance of the information.” Id. at *3. It noted that, the CDN logs were not “in any way searchable or [able to] be readily parsed.” Id. at *3. Moreover, it was “more than plausible that the plaintiffs have other ways of testing the veracity of PicClick’s assertions on this front such as, for instance, getting discovery into PicClick’s servers.” Id. (emphasis original).

As Lickerish illustrates, when discovery involves dynamic data, it pays to be prepared and to affirmatively raise the burdens of preserving, collecting, and producing dynamic data early in discovery.