The volume of documents routinely subject to discovery poses challenges in investigations and litigation that extend beyond e-discovery. While predictive coding is gaining increased acceptance as a procedure for identifying responsive documents with less manual review, there is less appreciation of how document analytics can add value in answering document related research questions, or otherwise helping to identify and analyze documents in ways not practical with keywords alone. Having reduced reliance on manual document review to decide which documents to produce, the challenge is to determine quickly what the documents reveal about the critical issues in the case.
Document analytics offer large potential payoffs in the conduct of investigations and case development. An advantage of using computer programs (i.e., algorithms) to analyze documents is that, unlike manual review, algorithms can be run across all documents in the universe at relatively limited cost. While the results of computerized document classification may not be perfect, analyzing all documents collectively reveals patterns not visible from targeted manual review. For example, important patterns of communication concerning particular topics may only become apparent once all messages are analyzed and mapped. Furthermore, algorithms can be used to gather individual pieces of similar information of interest across an entire database, for example pricing information, providing a basis for economic analysis that would otherwise be far more cumbersome to perform.