Nov 7


Originally posted on October 31, 2012


Judicial Activism with Predictive Coding – All In The Name of “Just”, “Speedy” and “Inexpensive” Discovery

I had a spirited discussion on Friday with a litigator about whether it was appropriate for Honorable J. Travis Laster from the Delaware Chancery Court to push the parties to use predictive coding when apparently neither of the parties sought to use predictive coding in EORHB, Inc., et al v. HOA Holdings, LLC, C.A. No. 7409-VCL (Del. Ch. Oct. 15, 2012).  I decided to take some time to think about this outcome since I have been very public in my support of using predictive coding to reduce many of the ills of the litigation system, e.g. lengthy discovery disputes, excessive costs and uneven quality of ESI productions.  Yet I am also on record supporting the outcome of Kleen Products where the plaintiffs failed in forcing the defendants to use predictive coding because they had what appeared to be an uphill battle to impose their discovery will on their adversary.  I also generally believe parties should be able to chart their own course for how to proceed with discovery.

So what makes the Delaware case different from Kleen Products?  It appears that discovery is about to start in this case so, unlike Kleen Products, neither party has expended tremendous resources pursuing a discovery strategy.    In addition, the Rules of Civil Procedure strongly support this Judge’s suggested course of action.  The Delaware Chancery Courts’s rules of Civil Procedure state:

“Rule 1. Scope and purpose of Rules.  These Rules shall govern the procedure in the Court of Chancery of the State of Delaware with the exceptions stated in Rule 81. They shall be construed and administered to secure the just, speedy and inexpensive determination of every proceeding.”

This language mimics Rule 1 in the Federal Rules of Civil Procedure.  Rule 1 is important as it sets the tone for the entire set of Rules of Civil Procedure.  So in any federal court, as well as state courts which start their rules of civil procedure with a general structure focused on “just”, “speedy” and “inexpensive”, the use of, and experimentation with, advanced TAR techniques like predictive coding should be encouraged by both parties and judges when it is appropriate (not all collections are suitable for methods like predictive coding; for example, audio and CAD files do not cluster).

Judges are lawyers too and charged to neutrally “administer” the Rules of Civil Procedure and attempt to help resolve disputes fairly in accordance with these rules.  Taken broadly, it comes as no surprise that a judge might encourage parties to use predictive coding to secure a “just, speedy, and inexpensive” discovery outcome.  This is because predictive coding used correctly, promises to reduce costs and turn over at least as much responsive ESI and less unresponsive ESI than our current eyeballs-on-every-document approach.

I also have a strong caveat, however in reaching this conclusion.  I do have significant concerns if this result goes beyond encouraging sporadic experimentation it could be an unintended disaster if every judge ordered the use of predictive coding.  As someone who spends a good deal of time working with lawyers and educating them on using predictive coding, I can say there is not enough knowledge in most law firms on how to use these processes and tools without significant vendor support.  Using technology is seldom a smooth path in litigation and lawyers producing ESI need more experience in order to understand how to react to issues that will occur in predictive coding projects.  Review teams will need to know how to assess validation statistics, use common sense judgment, and be ready to show that they have provided a reasonable discovery outcome.  Yet there are also not enough vendor resources to handle every case in the country as most vendors do not have a deep bench of project managers and their technical experts are stretched thin covering several hundred cases around the country.   While judicial encouragement of predictive coding is great and absolutely necessary, blind encouragement could be dangerous.

A final thought about this case is the Judge in addition to ordering the use of predictive coding unless the parties object, also mandates that they chose a single vendor to host the data. This ‘go use predictive coding’ without a defined process is a recipe for unfair/unjust results. Who does the vendor really work for?  How can that vendor provide any real support without favoring one side over the other? In the majority of cases that I have seen, the vendor is playing a significant role in assisting the lawyers with assessing the predictive coding results.  So there is a real challenge in finding a neutral vendor who can help both sides.  Who pays for the vendor?  In this case, this hubbub over predictive coding is from one last second, ad hoc order, pages 66-67 with no room for rebuttal or disagreement.  There is an argument to be made that this is a well-intentioned, but possibly uneducated bench that is forcing parties to use and pay for an undefined, black box marketing label.

The encouragement of experimentation with predictive coding was the exact conclusion of the Rand Corporation’s Study on predictive coding by Nicholas M. Pace and Laura Zakaras. Where the Money Goes: Understanding Litigant Expenditures for Producing Electronic Discovery Santa Monica, CA: RAND Corporation, 2012.  This study, after building a compelling case for the impossible challenge of handling growing volumes of ESI with traditional review and eyeballs on every document, concluded that the only way to solve these issues is by having lawyers experiment with more technology solutions like predictive coding.  That same type of experimentation is what Judge Laster requested the parties to do in discovery in his case.  I have no problem with this outcome as a single case or in a handful of other instances if other judges are interested in fostering the use of more technology to reduce discovery burdens on parties and the questions I raised above about neutral vendors are answered.  To the parties involved, good luck using the tools or crafting an argument on why you shouldn’t have to do so.  I am sure many of the country’s litigators will be interested in watching how discovery proceeds in this case and whether its results are “just”, “speedy” and “inexpensive”.

eDiscoveryJournal Contributor Karl Schieneman

- See more at:

leave a reply

Recent Posts

Recent Comments