Originally posted on November 27, 2012
I have been writing a series of posts about the “ever-elusive” metrics that many eDiscovery professionals seem to be waiting for when it comes to driving mainstream adoption of TAR. Does the elusive challenge of finding and providing these TAR metrics mean we are doomed not to be able to use TAR? No, it means clearly that lawyers, who have that rare blend of statistical training and court room experience, have a huge advantage in arguing in discovery disputes. The fact is there are boatloads of metrics that a smart litigator can use to defend their process. One analogous observation is the joke people have about statisticians. That two statisticians can argue forever that they are right on opposite sides of many issues by using statistics. We are going to have these arguments given the fact that the richness of collections vary and disputes have different values so these metrics will be moving targets. But since lawyers lead this argument, it is easy to see that if you put a lawyer in a room who understands this subject matter, against a lawyer who doesn’t, guess who is going to sound more reasonable? This is often the underlying standard being aimed for in discovery today. The analogy I have been giving people is that this match up tends to make the math adverse litigator look very much like a person on their first date; very unsure of how things are going and also what they should do when it comes time to present an adverse argument.
My advice to corporations who want to do TAR, and the recent FTI White Paper study suggests that over 60% of in-house attorneys want to do this versus 30% of outside lawyers, is to find a lawyer who can do this effectively. Advice From Counsel: Can Predictive Coding Deliver On Its Promise? FTI Consulting, By Ari Kaplan and Joe Looby. That is where the best investment of time can be made by organizations who want to use TAR. In contrast, most organizations vet tools ad nauseam, but for me, it’s less about the tool and more about the lawyer who can discuss the tool and its approach intelligently. This is because even though tools differ and workflows using them differ, if the lawyer running the project cannot advocate for the tool selected, there is an pretty fair chance they will have problems making coherent arguments for their approach. This holds true in meet in confers,in front of the court or even when they used their selected tool in the most efficient manner.
Two good examples of this on opposite ends of the spectrum are the 600 pages of testimony in the Kleen Prods. LLC v. Packaging Corp. of Am. N.D. Ill., No. 10-cv-05711, complaint filed 9/9/10 case which never could seem to intelligently focus in on the important issues as lawyers could not elicit clear testimony from their experts to help them with their position. Perhaps this was just strategy to confuse the judge and all the interested parties reading the materials. It might also be examples of lawyers discussing topics which they are not comfortable with. Contrast that with the Global Aerospace Inc. v. Landow Aviation LP Va. Cir. Ct. (Loudon Cty.), Consolidated Case No. CL 61040, Order Approving the Use of Predictive Coding in Discovery entered 4/23/12 case where a Carnegie Mellon trained chemical engineer, who happens to be a first chair trial lawyer, argued that statistics and sampling were a more reasonable way to approach the 2 million document collection and conduct a relevance review. If you read the transcript or listen to the audio file of the argument, you will see it was a fairly one sided discussion as the technology trained litigator comfortably was able to use metrics and industry statistics to make a compelling argument for using TAR.
So the conclusion of this series of articles is that, of course, metrics are important in TAR and are not for weenies. The real weenies might be the parties who do not understand TAR is chock full of metrics and when they encounter a lawyer who does, guess who is going to end up having sand kicked in their face? Given the pace of change in the field, this might take a few years, but law firms’ litigation departments might want to control this risk by hiring a few litigators who are comfortably able to argue positions based on statistics and sampling.