The Myth of the Easy Button
With Judges accepting application of Predictive Analytics, why haven’t we seen a massive conversion to technology based solutions?
Since the introduction of analytics to the eDiscovery space, there has been an assumption that once there was judicial acceptance there would be a massive migration to the faster, cheaper solutions leveraging the new technologies. Studies including the infamous TREC one have consistently shown that human review is far from flawless and consistent. And yet, there has not been a mass exodus from linear human review to automated or semi automated solutions. Why is this the case?
What the Heck is TAR?
The resistance to fully embracing analytic solutions comes from several places, but one of the largest sources is an overall lack of understanding as to what predictive analytics can and cannot do. And a false belief that to use Technology Assisted Review solutions (TAR) requires abdication of all control over the review process. The reality is quite the contrary, properly applied, TAR brings the discovery process back into the active fact driven case development process and allows attorneys to manipulate and analyze information from the outset instead of waiting several weeks for a review to complete.
So, what exactly is TAR, CAR, Predictive Coding/Analytics and what does it entail? They are solutions along a continuum that use the application of technology to the data involved with an electronic discovery review to reduce the volume of data that needs to be manually reviewed by humans. This can be done in several ways: prioritizing batching to ensure that the most likely relevant information is reviewed first, advanced culling of data prior to loading it into a review tool, call suggesting, through to applying an algorithm to predictively code data once a model has reached a point of stability. Along the spectrum there are dozens of potential applications of technology to review and the options are varied. How analytics are applied to a data set determines the level of control and oversight that a practitioner can maintain. In many cases, the early access to data allows for greater control in shaping the case development. There can often be greater success in manipulating data with a trained algorithm than application of traditional key words.
Why isn’t Linear Review Dead?
Everywhere that one looks today in the eDiscovery space it seems that some form of TAR is the magic bullet of the moment. Why hasn’t this killed traditional linear review? The answer in my mind is: 1) linear review makes more sense to someone outside of the insular eDiscovery space and 2) no one wants to be the person who gets sanctioned for being too aggressive in the application of technology and 3) properly applied a key word based advanced linear review can be a good solution for certain situations.
The solution to all three concerns is to take a measured approach along the spectrum of technology leveraged solutions. TAR is not a black and white solution that is either applied 100% or not at all. Aspects of the various technologies on the market can be applied and incorporated into a workflow that is very similar to the linear one that attorneys and clients are more familiar with. Whether you use it as an ECA component, for prioritization or have a second layer of human review to validate sampling, the cost and time benefits to TAR can still be realized. Taking this measured approach also insulates a practitioner from being sanctioned so long as the PROCESS is defensible, repeatable and easily explained to the judge.
Cat Casey
Vice President, Managed Review
eDiscovery Cat 