Originally posted on December 18, 2012
On September 9, 2011 I presented to a blue ribbon panel of in-house lawyers, outside counsel and judges at the Committee Meeting on Preservation and Sanctions in the Western District of Pennsylvania who were in agreement with my thesis that we over preserve, collect and process ESI in proportion to what is eventually used. My views on this topic have been evolving over time and I outlined this for the panel with a contrast of the Rand Study, Where the Money Goes with the Microsoft Letter dated August 29, 2011.
The Rand Study reported that, for the seven corporate participants, collection costs were estimated to be only $.08 of every dollar spent on eDiscovery, with the bulk of the expenses going to processing, and review. The studies’ authors and I tended to diminish the importance of these more minor collection costs and I believed we needed to focus on search and retrieval to lower eDiscovery financial burdens. But my initial conclusion failed to take into account all the silos of data being preserved under legal hold. I realized that legal hold preservation is much more frequent than active litigation and the Microsoft Letter supported this conclusion with data showing the vast over preservation of the average Microsoft matter vs. the actual amount produced and then used in litigation.
Microsoft reported that their average case starts with 48,431,250 pages preserved. Of that preservation, 12,915,000 pages are processed, 645,750 pages are reviewed, 141,450 pages are produced and 142 pages are actually used (Microsoft Letter, Page 5). This scary break down highlights the incredibly wasteful effort in handling a vast majority of ESI that isn’t even remotely relevant to the active matter. Additionally, from the predictive coding standpoint, when the percentage of responsive documents in the preservation and collection is small, it is logically harder to find responsive documents regardless of whether a predictive coding application uses random sampling or a seed set for training.
This is such a systemic problem that it really calls for outside intervention so I’ve decided to use Rule 1 from the Federal Rules of Civil Procedure as my letter to Santa.
This year, I would like “Just”, “Speedy” and “Inexpensive” eDiscovery which Rule 1 offers as a framework for interpretation of the entire Federal Rules of Civil Procedure. These concepts are even more basic than “reasonableness” and “proportionality” which eDiscovery lawyers often point out as the more important standards because they are inferred specifically in the eDiscovery rules in the Federal Rules of Civil Procedure. This would be acceptable, as long as “just”, “speedy” and “inexpensive” are in your sack of presents because they are easier concepts for lay people to grasp and they frame how “reasonableness” and “proportionality” are decided.
Santa, we’re obviously in the perfect eDiscovery storm with this blizzard of preservation and over collection so I need you to bring in your secret weapon, Rudolph! We need him at the head of the sleigh guiding the other ESI reindeer: Keeper – the Records Manager, Geeky – In-house IT, Traveler – in-house eDiscovery, Worried – outside counsel, Manager – In-house attorney, and Deeply Concerned the Client. Only Rudolph can shine his bright light on the process and satisfy any judge who turns into the abominable snowman and sanctions the team for poor preservation or spoliation.
Rudolph will help deliver a “speedy” and “inexpensive” solution by focusing, initially, on the areas where responsive ESI is most likely. This targeted collection can be fully analyzed during processing and TAR and if analytics show this data leads to other custodians, more data can be collected and become part of the evolving TAR population. Rudolph’s “speedy” and “inexpensive” sleigh ride brings us the third Rule 1 requirement: “Just”; because when litigation teams find ESI more quickly, they still have time to follow additional trails of information: new custodians, facts and theories about the case. Rudolph’s laser-beam approach means parties can begin producing ESI on a rolling basis, much earlier in discovery and performed supplemental searches for more ESI can be conducted if necessary. This method even follows much of Sedona’s Best Search Practices via “iterative” predictive coding.
Santa, we all know that Rudolph is real because we’ve seen his hoof prints in the snow of several recent cases. In Global Aerospace v. Landow Aviation (the “Virginia Case”) a more “just”, “speedy” and “inexpensive” result was derived cooperatively among opposing parties, and ultimately presented to the Court. During discovery, the selected software was trained by coding only 5000 documents, which were used to cull nearly 90% of the roughly 1.3 million documents without expensive and time-consuming manual review. This made Global Aerospace the first and only predictive coding case to be effectively implemented and concluded under judicial scrutiny in an active litigation. We can also nominate Vice Chancellor, J. Travis Laster, for the “Rudolph eDiscovery Award” for ordering both sides to use TAR and also, to share the same vendor in EORHB, Inc., et al v. HOA Holdings, LLC, C.A. No. 7409-VCL (Del. Ch. Oct. 15, 2012), known as (the “Delaware Case”).
Getting back to over preservation and collection, we need Rudolph to bring the sleigh earlier in the litigation process to have a more powerful impact on the Discovery burden than just leveraging predictive coding during review. We need one of the eDiscovery reindeer to don the red nose and ask the questions: why are we doing this so expensively? Why are we preserving so much data when we ultimately use so little? Reindeer Worried – outside counsel or Manager– In-house attorney would be a perfect choice for guiding the sleigh because they are lawyers and they understand the legal process. Deeply Concerned, Geeky, Traveler, or Keeper could definitely raise objections to over collection and preservation and even an outside consultant who is also a lawyer, call that reindeer, Thinker, can assist teams in deploying Rudolph’s laser-target method. It doesn’t take a computer scientist to realize the hole we dig ourselves into when we over process data for search and retrieval, even when we try to use technology to aid in the effort so maybe you could leave the North Pole a little early, this year.
Thank you, Santa.
Your ESI Pal,
With that in mind, I would like to wish everyone in the eDiscovery Community a happy and healthy holiday season and a Happy New Year to my good friends at the eDiscoveryJournal who have been a fun and resourceful group allowing me to conduct useful research with them and to let me blog to a growing audience of readers over the past six months. So let’s stop playing reindeer games and all put on our red noses to try to solve this problem. “Now, KEEPER! now, GEEKY! now, WORRIED! and MANAGER! On, THINKER! on, DEEPLY CONCERNED! on TRAVELER! and RUDOLPH!… Hopefully we have the makings of an everyday litigation classic in 2013 and beyond.