Archive | eDiscovery

The Emergence of Prediction in eDiscovery (LPO)

By Deepti Krishnan


2013 looks to be a slow growth year for Legal Process Outsourcing (LPO). 45% of the buy-side enterprises surveyed in the “State of Outsourcing 2013 Study” said that they had no plans to outsource or were going to decrease the scope of their LPO projects in the next 12 months. Only 19% plan to venture into a new LPO partnership or increase the scope of their ongoing LPO relationship.

LPO providers are also facing increased competition from new entrants that are reaping the benefits of onshoring and nearshoring. Carillion, a construction company, had a deal with CPA Global to outsource legal work. But now, it uses its own LPO venture: Carillion Advisory Services (CAS). CAS operates out of Newcastle in the United Kingdom, and is staffed by 60 paralegals. Another big entrant into the already crowded LPO provider space is Capita, an outsourcing giant, which recently added LPO to its service portfolio. It signed its first major deal with Pinsents in 2011. Capita’s 550-seat LPO center is located in Poland.

While law firms and corporate counsels are still outsourcing legal processes, the type of work is low-value. In Association of Corporate Counsel’s 2013 survey of Chief Legal Officers, eDiscovery is one function, which is primarily outsourced. This process forms the bulk of legal work outsourced, however, most eDiscovery firms are not traditional LPOs. For instance, Kroll, an investigation firm, has a suite of products, which includes the eDiscovery software “Verve”.

Discovery costs are directly proportional to the number of documents, where typically clients are charged USD 1 per document. If a case has 5 million documents for review, then the client will have to pony up USD 5 mn. eDiscovery technology that incorporates predictive analytics brings down the cost significantly, thus increasing the technology’s attractiveness to law firms and corporations with tight budgets.

The method by which predictive coding is used to determine whether a document is relevant to the case at hand is by developing algorithms to look for specific phrases and text in the document. For the legal industry, predictive coding enables a program to predict whether documents should be classified as responsive or non-responsive to a discovery request, relying on input by attorney reviewers on a sample set. Predictive coding reduces the cost of eDiscovery as the pricing is no longer linear. Most of the expense is incurred when grading the sample set and the cost increases marginally with an increase in the size of the document set. A 2011 study, “Technology-Assisted Review in E-discovery Can Be More Effective and More Efficient than Exhaustive Manual Review”, found that technology-assisted review can achieve at least as high recall as manual review, and higher precision, at a fraction of the review effort, and hence, a fraction of the cost.

Vice Chancellor J. Travis Laster of the Delaware Court of Chancery recently ruled that predictive analytics should be used in document review. This will help reduce costs as time taken to review legal documents for relevance will decrease. However, predictive coding only works in cases where the documents are text-heavy, not in patent review where documents have graphics. It also does not work in needle in the haystack investigations, as predictive coding is in its infancy, and privileged documents may slip through the cracks.

LPO providers that offer eDiscovery solutions need to incorporate predictive coding into their technology. Not doing so might sound the death knell of this line of business as new eDiscovery software with prediction algorithms come on the market. eDiscovery customers will migrate to the provider that offers the most cost-effective solution and adoption of predictive coding will ensure lower costs for both parties.

- Deepti Krishan, Research Analyst, Value Notes

Posted in eDiscovery, Industry Reports, LPOComments (0)

Page 1 of 11

Speaking English in the Philippines 2

Our Strategic Partners