Blog

Switching Your E-Discovery Horse Midstream – Can You Jump on a Predictive Coding Pony?

What if your discovery method wasn’t working out? Could you make the jump?

In the last few months, two different federal courts considered similar discovery scenarios, but reached opposite conclusions about how discovery should proceed. The scenarios looked like this: The parties agreed to a discovery plan that did not include the use of predictive coding. But after wading into discovery, the producing party realized that, in light of the volume of documents to be reviewed and the burden of conducting review with traditional methods, predicting coding could be worthwhile. The producing party sought agreement from the opposing side to change approaches and switch to a predictive coding workflow. The other side refused to agree, however, so the producing party asked the court to resolve the dispute and approve the switch.

Should the producing party be permitted to make this switch?

Based on these two recent decisions, the answer rests on the level of transparency and cooperation it is willing to adopt.

In Bridgestone Americas, Inc. v. International Business Machines Corp., No. 3:13-1196 (M.D. Tenn. July 22, 2014) (Mag. J. Brown), faced with a review of more than two million documents that hit on an agreed-upon set of search terms, producing party Bridgestone sought to use predictive coding to identify responsive documents for production. Defendant IBM opposed the request, objecting that it was an unwarranted change in the original case management order, and asserting that it is improper to use predictive coding after an initial cull using search terms. (For a discussion of two other cases that addressed the search-terms-then-predictive-coding approach, see my blog post, Predictive Coding and Freedom of Choice: Sedona Principle 6 in Action, regarding the Biomet and Fosamax decisions.)

Magistrate Judge Brown observed that the use of predictive coding is “a judgment call,” and “there is no single, simple, correct solution possible under these circumstances.” Reminding the parties that discovery should be tailored to be as efficient and cost-effective as possible, the judge decided to permit Bridgestone to use predictive coding on the documents found by search terms. The judge acknowledged that he was “allowing Plaintiff to switch horses in midstream.” But he admonished the parties that full “openness and transparency” will be critical, and set an expectation that the parties and their experts will communicate frequently.

Just two months earlier, in Progressive Casualty Insurance Co. v. Delaney, No. 2:11-cv-00678-LRH-PAL (D. Nev. May 19, 2014) (Mag. J. Leen), Progressive asked the court to amend the parties’ stipulated ESI protocol and allow the use of predictive coding on documents returned by search terms. In this case, however, Magistrate Judge Leen rejected the request and disallowed the switch in approach. Like Judge Brown, Judge Leen expressed support for the use of predictive coding as a tool that can streamline discovery and make document review more accurate and effective. But under the circumstances, the court found that it would not be appropriate to permit Progressive to abandon manual review and use predictive coding instead.

The court pointed to several factors that influenced its decision:

  • Progressive began using predictive coding to review ESI without the opposing side’s agreement to amend the stipulated ESI protocol, and without seeking leave of the court.
  • Even after involving the other side, Progressive was “unwilling to engage in the type of cooperation and transparency that its own e-discovery consultant has so comprehensibly and persuasively explained is needed for a predictive coding protocol to be accepted by the court or opposing counsel as a reasonable method to search for and produce responsive ESI.”
  • Progressive’s proposed protocol failed to follow best practices recommended by Equivio, the developer of the predictive coding tool it selected.
  • The switch to a predictive coding workflow would further delay an already-delayed discovery schedule.

Based on these considerations, the Magistrate Judge refused to permit Progressive its requested “do-over” of discovery using predictive coding. These two decisions illustrate the importance to courts of cooperation and communication when a party seeks to use predictive coding to accomplish a document production. Whether at the outset of the matter when the parties are beginning their discussions about a discovery protocol, or after discovery is underway and one of the parties wants to switch to a different approach, litigants should come together and work collaboratively to resolve issues around the implementation of predictive coding. (Or if a party chooses to use predictive coding without transparency, it must be prepared to defend its methods to the court and assume the risk of being forced to re-produce documents using alternate means.) Whatever the chosen approach to cooperation, a producing party should engage qualified experts to help develop and implement its protocol.

DiscoverReady and its experienced team of eDiscovery consultants understand how to leverage predictive coding and other advanced analytics to reduce the overall cost and burden of discovery. We hope that more litigants will turn to these tools in the right circumstances, even if it means asking the court to approve a midstream change of horses.

Maureen O'Neill