Situation: 1.7 M documents quickly reduced to 25,000 despite unfavorable search terms.
A dispute over licensing fees forced a digital entertainment service into litigation, unfortunately for the second time in as many years. The entertainment provider did not want to relive the ordeal it went through during the first document discovery process, when it employed a time-consuming and largely manual review of the company’s documentation. This resulted in a surfeit of largely irrelevant data and a cost that surged well past initial projections.
Facing the daunting task of evaluating 1.7 million documents, the company turned to FTI Technology, the e-discovery practice of FTI Consulting, to employ its analytics-based e-discovery services. FTI was immediately confronted with a major challenge: the attorneys for the opposing litigants refused to renegotiate the previously agreed-upon keywords that could be used in document searches for case-relevant material. Unfortunately, the keywords FTI needed to use were generic and ubiquitous—making it unusually difficult to eliminate the irrelevant documents. Adding to the pressure was a narrow three-week time window to finish the job, as well as a demand by the opposition to receive no more than 30,000 responsive documents.
In order to avoid giving the plaintiff in the case cause to challenge the validity of the disclosed documents, it was crucial for FTI to develop a strategy and customized workflow that suited its client’s unique needs.
FTI used its Ringtail e-discovery software solution to drastically reduce the number of documents requiring further examination. Even though the established search keywords were overly broad, FTI still hoped that predictive coding in and of itself could be an effective solution. From a time-savings standpoint, it seemed ideal to perform just the one process.
However, an initial sampling of documents resulted in an overabundance of responsive documents. Ultimately, it was determined that predictive coding alone would have necessitated a thorough review of 360,000 documents—far too many for the client’s needs.
In response, FTI developed a specialized workflow for its client by using Ringtail’s Mines and Cubes data mining tools to organize the documents into various themed subsets. These themes were then cross-referenced with the coding from FTI’s predictive model testing, in order to determine which themed groups most closely correlated with findings of relevance or non-relevance. Themed groups discovered to be highly relevant were immediately reviewed, while those with no relevance were quickly discarded. The remaining themed groups were then subjected to a new, more effective round of predictive coding.
Leveraging the functionality of Mines and Cubes, FTI initially eliminated almost 985,000 of the 1.7 million possible documents from consideration. A total of 20,000 docs were identified as high priority and immediately reviewed, ultimately yielding a precision (or relevance) rate of 72.8%. The remainder underwent a second round of predictive coding, which resulted in the removal of an additional 600,000 documents.
Rather than reviewing 360,000 docs, as would have been necessary had just predictive coding been applied,the entertainment provider now only needed to scrutinize 59,000 docs, of which 25,000 were ultimately sent to opposing counsel. Attorneys for the opposition issued no discovery challenges.
Impact: Efficiencies resulted in a cost savings of $2.5 million. Nearly a million irrelevant docs were promptly eliminated
The efficiencies generated by FTI’s services represented a savings of $2.5 million over a traditional linear review of the same documentation. Moreover, the revised, optimized workflow process created specifically for this case generated a savings of $600,000, compared to a scenario where only the predictive coding tool was used.
This initiative was so successful that the digital entertainment service contracted FTI for additional legal matters. The company has already saved $123,000 in up-front costs by reusing the same repository of information from the previous litigation.