Defensible Predictive Coding Will Change The Economics Of eDiscovery In 2010

Tuesday, March 2, 2010 - 01:00

Editor: Mr. Robman, would you tell our readers some background on yourself and Recommind?

Robman: First, thank you having me. I am Assistant General Counsel at Recommind and oversee the company's daily legal operations. I have been actively involved with eDiscovery for over 10 years now in various capacities. My first taste of discovery was as a young associate, back when we conducted document review and coding with a clipboard and dust mask, when most of the records were paper and stored in large warehouses. I then moved in-house at Bank of America, where I was part of the Regulatory Investigations Group. From there, I moved into the consulting sector, where I focused on complex litigation matters, securities litigation, antitrust matters, internal corporate investigations and electronic data analysis and computer forensics. Basically, I have seen eDiscovery issues from all sides (well, except as a government regulator).

Recommind is the leading provider of search-powered information risk management software. Specifically, we provide our clients with enterprise search, eDiscovery and email management software. Our customers include law firms, corporations and government agencies both domestically and abroad. We are headquartered in San Francisco with offices worldwide.

Editor: Many of our readers are concerned with escalating eDiscovery costs. What are some of the strategies they can implement to reduce cost?

Robman: Companies should be concerned about the rising costs of eDiscovery; as the amount of data continues to escalate, so do the risks. The best advice that I can give to proactively manage corporate data is: (i) know what you have and why you are keeping it, and (ii) conversely, know what you no longer have and why you deleted it. Of course, a company must maintain compliance with its legal and regulatory data retention obligations, but by proactively managing data, the company can reduce the amount of data that is potentially subject to review, preservation and production in litigation and regulatory scrutiny.

In terms of cost reduction strategies, I would suggest a spend analysis to determine how much is coming from internal resources and, more importantly, from external sources, such as law firms, consultants and technology providers. I would then build a process map for how you currently handle eDiscovery matters to find areas where you can automate the process and find cost efficiencies. For example, is your default process to hire consultants to forensically image hard drives for every matter? Or, should the default process be to conduct strategic in-place review to determine what and whose data is potentially responsive on each matter? The latter strategy allows you to save costs, minimizes the impact on your employees and reduces the amount of data being preserved.

Once you have your process in place, I would then head to the left of the EDRM model towards the Information Management phase, where real proactive information management can occur to reduce the overall amount of ESI being stored. Categorization and email management technologies are mature enough to automate the classification process early in the ESI life cycle. Thus, a company can now leverage automation within its process to understand its ESI. I would be looking for technology that identifies distinct concepts within data, regardless of the language or subject, and organizes and prioritizes data accordingly.I would also evaluate the technology with my own company's data to ensure that it delivers automation at a high level of accuracy.

Editor: We've been hearing that new technology is part of the solution to the voluminous amounts of data being subject to review, but is it a panacea?

Robman: I will answer that question in terms of two technology advances in eDiscovery that will have a dramatic effect on reducing data volumes. First, technology alone cannot remediate the problem, but technology coupled with the right documented, defensible and monitored process/workflow can make a huge difference in reducing the costs associated with legal review. By now, many have heard of Defensible Predictive Coding, which Recommind has been talking about for some time: it is a dramatic shift from the traditional linear review process. The uses of Predictive Coding are case dependent; however, we are seeing clients save 50 to 80 percent of what they would have normally spent on traditional review because contract attorneys were either not needed or were able to work far more efficiently (or both). I think it is a not-so-well-kept secret that linear review alone is typically not an overly accurate or consistent review methodology, but it has been the devil we know. Predictive Coding is a quantum leap in document review efficacy and accuracy.

Next, we are in the midst of technology advances in the area of legal holds and early case assessment (ECA). Legal counsel is being provided with tools to gain a better understanding of the case much earlier in the case life cycle and technology is playing a role. Technology that allows users (such as legal professionals) to conduct in-place review (or what we call Explore-in-Place) enables them to find, sample and assess ESI before collection provides real ECA, not just culling of post-collection data. That allows companies, number one, to conduct ECA as early as possible - which is the whole point with ECA - and number two, only to collect the ESI they need to collect for any given matter.This far more targeted and precise approach in turn significantly lowers eDiscovery costs and makes them identifiable and predictable from the very outset of a matter.

Editor: Is the sole focus on reducing costs or are there other factors that they need to be aware of?

Robman: Certainly cost is a large factor, especially in this economy. But, those companies that can manage and decrease their risks are the ones that will thrive. Of course risk comes in many flavors - legal, compliance, operational or productivity. Major litigation or regulatory investigations have a dramatic effect on a company and can grind productivity to a standstill. Sometimes litigation or regulatory scrutiny is inevitable, but those companies that have a documented process in place are able to reduce the impact. There is a sense of calm that comes with a documented process and tested technology during the course of litigation. In contrast, there is a sense of panic when litigation strikes and the company has no response plan. Panic leads to mistakes and a reliance on third parties, which will only cost you more in the end.

Editor: What other types of risks should a company be concerned with?

Robman: Avoiding reputational harm, or at least keeping it to a minimum, must be a major consideration. Most companies devote significant resources to building customer loyalty or public trust, which can be eroded very quickly. Part of reducing the potential reputational harm is having an early understanding of the case so you can evaluate whether to settle or proceed. Part and parcel of that assessment is the ability to understand the costs that will be required to defend the case; the strengths and weaknesses of your case; those potentially involved; and the impact upon your business. A company that has the process and tools to conduct this type of assessment up front can truly manage and diffuse reputational harm.

Editor: What are some current cases that our readers should be aware of?

Robman: I think that if readers read one case so far this year, then it should be a case from the middle of January when U.S. District Court Judge Shira Scheindlin of the Southern District of New York (yes, the same judge of Zubulake fame) issued an opinion in Pension Committee of the University of Montreal Pension Plan, et al. v. Banc of America Securities, LLC, et al., 05 Civ. 9016, 2010 U.S. Dist. LEXIS 1839 (S.D.N.Y. Jan. 11, 2010). This is the first major eDiscovery opinion in 2010 and likely will be the most remembered when we arrive at the end of 2010.

In short, Judge Scheindlin's opinion details the data preservation efforts of the numerous plaintiff investors who joined an action to recover an alleged half-billion dollars in losses from the liquidation of two British Virgin Islands-based hedge funds. In particular, the opinion details the dearth of plaintiffs' preservation efforts and ultimately held that that seven of the plaintiffs acted negligently and that six of the plaintiffs acted with gross negligence. The result of these preservation failures was the probable loss or destruction of relevant data requiring further discovery, monetary sanctions, and a carefully crafted spoliation instruction to the jury.

Editor: What could the sanctioned plaintiffs have done differently to avoid the results?

Robman: There are definitely some lessons learned in the Pension Committee case, although many thought those lessons had been the product of Zubulake a few years ago. Here are a few lessons learned:

Once the duty to preserve is triggered, counsel should (i) suspend auto-deletion of ESI; (ii) ensure preservation efforts take departed employees into consideration; and (iii) take reasonable steps to ensure preservation of relevant ESI under the litigating party's control and in the custody of third parties.

When a party reasonably anticipates litigation, it must suspend its routine document retention/destruction policy and put in place a "litigation hold" (also known as a "legal hold") to secure the preservation of relevant documents.

Consider the defensibility and risks posed by custodian self-collection. While it can be defensible with the appropriate monitors and supervision, it is important to understand the technology and process used for collection. Most importantly, action and decision points should be documented early and often.

I think we will be hearing much more commentary about this case and others on the duty to preserve relevant ESI.

Editor: What advice do you give to those who are starting from scratch?

Robman: Advice that I offer to everyone who is just starting out, whether you are a small company or large corporation, is that you should not put off starting the process in favor of trying to do it all at once at a later date. Nor should your initial goal be perfection. Finally, delaying because you think that your company will never face major litigation or regulatory scrutiny is not advisable.

The first step in the process after assembling the correct team is to conduct a risk assessment to determine what data sources, business units, products, etc. pose the most litigation or regulatory risk. I would target those areas of risk first and foremost. For example, many companies choose to start with email and derive a process to preserve, collect and review when called upon. The next step is to conduct a gap analysis on the process side. For example, are there are any existing data retention/destruction policies in effect and are they being followed? Where there are policy gaps, they should be filled. From there, move onto determining whether there are any technologies that can assist with your action plan. Finally, make sure to document your process and perform testing to ensure readiness.

Editor: Who needs to be involved in the process?

Robman: That is a great question to really wrap things up. Trying to solve ESI management and eDiscovery issues in a silo never works. You will need a multi-disciplined group from all corners of a company including representatives from Legal, Compliance, IT, Operations and the business units. In addition, there will need to be a champion who can bring process and technology together into a business case for senior management support. Top-down support for this endeavor will be crucial. Finally, look to the outside to fill the expertise and technology gaps such as your outside legal counsel and eDiscovery technology companies. Go forward with those who will act as your partner in the overall process, not just a single-step provider.

Please email the interviewee at with questions about this interview.