Break It Down: Identifying eDiscovery Costs by Task
It is well established that ediscovery costs continue to rise, and Gartner, Inc. estimates the market to grow to $1.5 Billion by 2013 (“Magic Quadrant for E-Discovery Software” by Debra Logan, John Bace, May 13, 2011). A recent study by the Rand Institute for Civil Justice, “Where the Money Goes: Understanding Litigant Expenditures for Producing Electronic Discovery,” addressed the cost of producing electronic documents.
Rand interviewed eight very large companies about ediscovery production expenses. Participants spanned the communications, energy, insurance, biotech and pharmaceutical fields. Case data from 57 large-volume cases was used in the study, and costs were broken down into three tasks: Collection, Processing and Review. As seen below, the largest expenditure for cases was the Review task at 73 percent.
The study discussed ways companies can attempt to reduce the costs of traditional review. With ediscovery attorney rates ranging from $450 for partners, to $250 for associates and $75 for contract review attorneys, it is difficult to reduce the per hour review cost. Moving document review offshore can be appealing on a purely cost savings analysis, but issues like quality, information security and confidentiality make companies reluctant to adopt this low-cost solution.
Rand identifies other methods for reducing Review costs, such as document grouping and computer-assisted review to speed up the process. The three most commonly used techniques for grouping documents are near-duplicate detection, clustering (a.k.a. topic clustering or concept clustering) and email threading. However, according to the study, the cost savings are not significant.
To significantly reduce the cost of review, the report suggests that companies need to move away from their reliance on “eyes-on” review. Rand states that computer-categorized review, which is commonly referred to as predictive coding, will likely result in significant cost savings for Review. It works by having the system automatically assign a rating or proximity score to a document that reflects how close the content is to another document that was coded by an expert reviewer. Predictive coding can be more accurate than traditional eyes-on review; however, it has yet to be adopted by the courts.
Rand acknowledges that there are still a number of barriers to mainstream computer-categorized review, including the obvious concerns around recall and precision. However, as ediscovery costs continue to soar, organizations will demand technology to help control them – a handful of technology vendors have already answered the call. These vendors are leading the way by delivering tools that can be used and understood by legal users. The industry’s ability to educate the legal community on the benefits and risks of the technology, and how those risks can be mitigated, are fundamental to the adoption of predictive coding. I anxiously await new developments in predictive coding and will continue to keep our clients abreast of happenings in this realm.