Roundtable: E-Discovery - An Ever-Improving, Revolutionary Development

Monday, August 30, 2010 - 01:00

Editor: Has the increased accuracy and efficiency of e-discovery software significantly reduced its costs as well as attorney costs in ferreting out significant documents?

Sharp: Technology is now available to identify significant documents. This is a sea change in the industry. The first generation of e-discovery technology has been able to seek out non-significant documents. At the basic level, this includes de-duplication and culling by file type, data range or custodian. More advanced examples of technology in this stage include near-duplication algorithms, and email threading, where content analysis is used to identify emails that are contained in subsequent layers of a thread. But once you've done all that, you still have the challenge of finding the significant or relevant documents among what's left. That's costly, often prohibitively so.

The new generation of e-discovery technology includes the ability to identify these significant documents. Initial attempts to address this problem were naïve applications of keyword searching, which is plagued by accuracy issues, or clustering by topic, which is only a tangential solution, organizing the documents but essentially leaving the heart of the problem - finding what's relevant - intact.

A new wave of technologies has emerged with the arrival of software that can learn from a sample set of coded records. The decisions of document relevancy from the sample are then propagated to the rest of the documents in the set. This allows litigators to review fewer documents, and review them in order of relevancy. We regularly see cost savings of over 50 percent through the application of learning technology.

Kibbe: Yes, in many ways the evolution of technology in the e-discovery space has resulted in real cost savings for the ultimate corporate client on both the e-discovery provider and outside counsel front. The real truth is that, in most cases, less than 20 percent of most document sets are relevant. However, the cost of identifying the subset of these relevant documents is extremely high, particularly when an attorney review team utilizes a linear review.

Today's advances in technology allow for the elimination of the waste associated with traditional linear review by replacing it with strategies that prioritize relevant documents and allow related documents to be reviewed as a group. Simply utilizing these two strategies can reduce overall attorney review spend by at least 50 percent while maintaining full defensibility, since every document is reviewed.

For example, today's prioritization technologies can be combined with good project management and workflows to identify the 20 percent population of the document set that is likely to be critical to the case. Once identified, those documents will be sourced to outside counsel for final privilege review, thereby eliminating the need for a first pass review. In addition to providing a significant cost savings, prioritizing evidence also allows for a true early case assessment of the risks involved in the case. Decisions to proceed and/or settle can now be made in days or weeks rather than months or years by having the ability to focus on the important documents early.

Prounis: Yes and no. Today's e-discovery solutions must always pay homage to the simple things, the blocking and tackling steps of e-discovery. By that, we mean the things that determine success or failure of the effort, i.e., things that allow you to declare an effort as having been done correctly or advise a regulator that a production was thorough and complete enough. Many of the tools on the market today only cover one portion of the waterfront and do little or nothing else. Usually, you need to have a sound and solid foundation in place to benefit fully from process enhancements, such as conceptual search, threading or near duplicate analysis. When properly deployed, these enhancements offer major quality and productivity gains. We deploy them for early case assessment, foreign language detection, "smart batching" during the review phase, enhanced privilege analysis and to ensure consistency of redactions. Given the rapid increase in ESI volume, new and better methods are always going to be required. They can be utilized without sacrificing any control, completeness and/or project efficiency! Absent quality control checks and critical system components, however, some misdirected E-Discovery products are less reliable and potentially offer partial and incomplete solutions. This can actually increase total project cost without ever realizing any true cost savings.

Some leading ESI processing tools actually ignore encryption, passwords, embedded files within documents, comments, notes or calendar items. Some e-discovery tools modify the metadata of the files upon export and effectively sever the chain of custody.

Some review tools don't have any repository component or production engines or efficient method to export data. You may be able to find something in an eloquent way, but either it takes forever to export the results or you have to reprocess the ESI, as it wasn't designed to go all the way to production. Some concept searching tools require weeks of software training, subject matter experts and concept weighting before they even can be used on a single file. What obligation do you have to assure yourself, your opponent and the court that the discovery performed by your side was credible and that your tools, process and quality controls work as promised? Most judges believe you have a solemn obligation to know these things! Only through the process of e-discovery standardization within your corporation or law firm will cost control and savings be achieved. Only then will you realize the benefits envisioned by your question. Partial tools, however good at doing one particular task, only result in partial solutions, which in 2010 simply will not cut it!

Gaines: With the advent of quantum leaps in data that is now discoverable, the answer is: it depends on the case. FRCP Rule 26(b)(2)(B) states "A party need not provide discovery of electronically stored information from sources that the party identifies as not reasonably accessible because of undue burden or cost."

Thanks to increasing technological dexterity, the definition of "reasonably accessible" has expanded to include vast seas of data. Prior to the advent of e-discovery capabilities, getting data off backup tapes was considered unreasonable in all but very "big" cases. It was cumbersome, time-consuming and required lots of human intervention.

Technology has now made data on tapes "reasonably accessible" in all but "small" cases. Indexing software has made analyzing terabytes of data "reasonable" in most cases.

These capabilities have an upside and a potential downside. The upside is that far more information is easily discoverable and available. More sophisticated analytics can result in more sophisticated search algorithms, which can find relevant information that would otherwise not surface. The added dimensions are that the bar has been raised on what is reasonably accessible and could result in more custodians and data locations or sources being added to the search universe. Further, indexing content and word context analysis can find correlations that a team of attorneys could miss. Also, social network analysis, now made economically possible with analytics software, can result in finding additional relevant custodians and records.

The bottom line is that e-discovery software can result in greater success in accessing and locating "reasonably accessible" data as well as run the risk of data spoliation and finding smoking guns. This might increase the risk of sanctions and/or an unfavorable decision in a case.

Whether more accurate and efficient e-discovery software increases or decreases the cost of discovery (or the overall cost of a case) is dependent upon several factors, including the e-discovery skills of the litigants and the judge, the discoverable records, and the factors in the case. There is no clear cut answer - it depends on the case.

Cohen: While it's clear there have been significant investments made in e-discovery software over the past several years, what's not clear is how much cost savings these applications have delivered in aggregate . That's because technology advances have been outpaced by the sheer volume and complexity of the cases requiring electronically stored information (ESI). For instance, many more cases require e-discovery today than even 24 months ago, and each of those require an increasing amount of volume - as email is no longer the only data type being targeted.

There's also an overriding question: has software actually increased the accuracy and efficiency of e-discovery? There is no question that technology like clustering has expedited the process, however, what we're hearing from clients is that there is increasing push-back from the courts on many of these advanced technologies because it's so difficult to prove that the search results are accurate due to the complex mathematics that underpin them.

The smart way to drive cost savings into e-discovery is to leverage the tried and true method of developing keywords to cull the data based on what the documents are about, and to build efficiencies into how those keywords are selected in the first place, thus streamlining every subsequent step.

Editor: Do you believe that e-discovery is on the brink of taking on the role currently held by client attorney? What new advances are being made?

Cohen: e-discovery software will never take on the role of the client attorney, nor should it. Such an action would require an unacceptable reliance on artificial intelligence to make critical decisions. Consider this: would you be willing to let a robot perform major surgery on you? Of course you wouldn't. The mind of a seasoned surgeon is far more adaptable to unforeseen circumstances than that of a computer. The same reasoning can be applied to a legal matter.

Technology has a very clear play in e-discovery, but its role is to supplement the human mind, not to replace it. Two examples of this type of technology from RenewData are Anagram & Vestigate. When combined, they enable the attorney to make intelligent, swift decisions based on the content of the case - striking the right balance between speed and leaving the heavy lifting to humans.

Sharp: On the contrary , we believe that this new "learning" technology actually empowers the attorney. The learning process draws on the input of an experienced, knowledgeable attorney, amplifying the knowledge and experience of the most savvy resource on the case. Moreover, the review function remains in place, only its scope is reduced. We refer to the "read less, think more" phenomenon enabled by learning technology. In recent years, the litigation attorney's focus has shifted to the soul-destroying, tedious role of document review. Learning technology can restore the litigation attorney to his/her traditional role as legal analyst and strategist.

Kibbe: We do not believe that e-discovery tools and protocols will ever supplant the critically important, strategic role that outside counsel provides to corporate clients. Rather, we view the evolution of electronic discovery technology as the means by which outside counsel can provide even better and faster advice to their clients regarding the risks involved in a litigation or investigation, thereby delivering better outcomes.