As seen by this survey result, the COVID-19 pandemic has certainly impacted organization budgets across the board. At the same time, we have already started to see increased litigation as a result of the pandemic. So, how does an organization manage an increased litigation workload on the same or less budget? Here are four best practices I ran across that I liked from process to machine learning about which I will provide my own thoughts and observations.
Brian Schrader, President & CEO of BIA, just published these tips in KM World (4 Practices to get your e-discovery process recession-ready), so below are his list of four practices with my comments:
Make eDiscovery a standard business process: If you read this blog regularly, you probably already know this one and have done so (or at least understand the importance of doing so). I particularly liked Brian’s comment here: “Think of the process as being similar to any other continuity plan you have in place. It should include step-by-step instructions for everything from initial legal holds to producing final documentation in court. The plan should be well documented and shared with key members of the organization so that each person understands their role.” Want a good book to read to help with that? Try this one from Mike Quartararo – the guy who literally wrote the book on project management in eDiscovery.
Consider remote collection and custodian questionnaires: This was already becoming very important before the pandemic – it’s now vital since the pandemic began. With eDiscovery professionals not looking to get back into the office anytime soon, the eDiscovery collection market may never be the same and so much more of it has started to be conducted remotely. One notable provider recently stated that less than 5% of their collections today are done on-site. So, it’s important to develop and continue to enhance questionnaires for remote collections from your custodians. They’re not coming back into the office anytime soon. Oh, and good questionnaires help assess custodians from a proportionality point of view as we discussed in this recent webinar.
Avoid the collection of unnecessary data: In that same webinar, we unveiled some Big Data stats that illustrate how much data organizations are dealing with in today’s world and Brian provided another one in his article – by 2025, it’s estimated that 463 exabytes of data will be created each day globally. How much is that? Over 463 million terabytes – per day!! Historically in eDiscovery, the data has moved to the technology, but more and more, the technology will have to move to the data to identify more precisely what’s potentially responsive and only move that forward. Good processes and better technology will help organizations keep those collection costs manageable.
Leverage machine learning for document review: Brian quotes the 2012 RAND study that states that 73% of all eDiscovery costs are spent on legal document review. It’s gone down, but not much – Rob Robinson’s Complex Discovery annual market Mashup from last year estimates “review-related software and services…to constitute approximately 69% of worldwide eDiscovery software and services spending in 2019”. It also noted that “[w]hile the percentage of spend is decreasing over time, the actual dollar spend is estimated to increase based on overall software and services market growth.”
So, review continues to be the most expensive phase in eDiscovery (by far), yet the 2018 ABA Legal Technology Survey Report (released in early 2019), stated that only “12% {of lawyers responding to the survey} report using predictive coding to process or review e-discovery materials”. Think, McFly, think! Predictive coding and machine learning technologies have been proven to provide results as good or better than manual review at a much more efficient and cost effective manner. If you want to reduce eDiscovery costs, start with the biggest one and use proven and court-accepted technology to do so. Makes sense, right?
Also, just a reminder that, on Wednesday, August 19th, HaystackID will conduct the webcast On the Case? eDiscovery Case Law Update for the First Half of 2020 at noon ET (11am CT, 9am PT). In this presentation, I will be covering key case law developments during the first half of 2020 – along with Ashish Prasad, Vazantha Meyers, Todd Haley and Seth Curt Schechtman of HaystackID – to identify important rulings that may impact how you conduct discovery going forward. Don’t miss it!
So, what do you think? Are there any other best practices you would suggest to manage eDiscovery costs in today’s pandemic influenced world? Please share any comments you might have or if you’d like to know more about a particular topic.
Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.
“If you want to reduce eDiscovery costs, start with the biggest one and use proven and court-accepted technology to do so.”
I would suggest that the biggest mindset shift that’ll help spur adoption is one not of absolutes, but of relatives. The terms “proven” and “court-accepted” make it sound like it’s a dangerous minefield, and if you step on the wrong technology, you’ll be in trouble.
The fact of the matter is that all supervised machine learning technology works.. though some technologies work better than others. A better mindset would be one of “relativism” rather than “absolutism”.. and being able to think about how one might compare — evaluate — one technology relative to another. (Aside: Evaluation is a topic I could spend pages and pages on.)
And then at the end of the day, understand that a strong argument can and should be made that “court accepted” does not refer to the technology, but to the outcome. Through statistical sampling, one can validate what one has done, no matter how one did it. And the outcome is what matters, not how you got there.
All these points, when taken together, should spur lawyers to do two things: (1) Get started immediately using supervised machine learning — any supervised machine learning — and not fear stepping on a landmine if they happen to pick the wrong one initially, and (2) Not be complacent in that first choice that they’ve picked. By knowing how to ask the right questions to do a proper relative evaluation, one should move quickly from that first choice to any number of additional options.. and actively be seeking out which consistently gives the best relative performance.
Great points, Dr. J, and I realize that I should have used the term “methodology” as opposed to “technology”.
What I was trying to make clear is that the use of computer assisted review methodologies such as supervised machine learning have been approved for use in courts since Judge Peck’s ruling in the Da Silva Moore case in 2012. Of course, any technology can fail if supplemented with a flawed approach (that’s why validation through statistical sampling is so important as you point out), but (as you know) supervised machine learning has been proven to work in many cases to effectively reduce costs with comparable or even improved accuracy. So, yes, get started using it, and learn and improve as you go (like everyone else does). 🙂
“What I was trying to make clear is that the use of computer assisted review methodologies such as supervised machine learning have been approved for use in courts since Judge Peck’s ruling in the Da Silva Moore case in 2012.”
100% agree. My comment came from a position of having heard over the years (not from you, but from more than a handful of others) that Peck’s ruling applied only to a specific technology, not to similar methodologies in general. So I just wanted to double clarify your clarity 🙂
I like the focus on remote collections. I tend to work even further upstream with clients to make sure that remote worker’s minimize creation of unique, offline ESI. The new generation of messaging/collaboration apps make that challenging at times. Most traditional ESI relevant to civil litigation without fraud or bad actor issues is synchronized into Office 365 and other corporate applications. Engineers, artists and other users working on very large files are a notable exception. It is better to have a copy of all the data for in place preservations and selective collections than have employee laptops remotely copying hundreds of GB over their home internet connection. Good commentary.