Because it will make your review more efficient and effective.
The process of searching and reviewing electronic information is typically the most expensive and time-consuming phase of any legal matter. Predictive Ranking can dramatically reduce the scope and time of a review—resulting in significant savings. Predictive Ranking achieves these reductions in two ways. First, it finds responsive documents far more accurately and efficiently than can be achieved using keyword search or linear review. Second, it orders documents by relevance so that your review team can start with the most important first.
The Weakness of Manual Review
Although manual, linear review was long considered to be the gold standard for accuracy in discovery, numerous studies have now established that technology-assisted review (TAR) is more accurate than manual review for finding responsive documents. One highly regarded study, conducted by Maura Grossman of Wachtell, Lipton, Rosen & Katz and Professor Gordon Cormack of the University of Waterloo, concluded, "[T]he myth that exhaustive manual review is the most effective—and therefore the most defensible—approach to document review is strongly refuted. TAR can (and does) yield more accurate results than exhaustive manual review, with much lower effort."
The Limits of Keyword Searching
Keyword searching is an important tool in e-discovery. However, it is also an imperfect one, undermined by imprecision in keyword selection and over- or under-inclusiveness in results. U.S. Magistrate Judge Andrew J. Peck, in his seminal decision authorizing technology assisted review, Da Silva Moore v. Publicis Groupe, cited the weaknesses of keyword searching, writing that "the way lawyers choose keywords is the equivalent of the child's game of Go Fish."
Another problem with keyword searching is that it often finds too few responsive documents or too many non-responsive ones. In the landmark e-discovery study by David C. Blair and M.E. Maron, An Evaluation of Retrieval Effectiveness for a Full-Text Document-Retrieval System, 28 COMMUNC'NS. OF THE ACM 289, 295 (1985), the attorneys in the study were confident that their searches had found more than 75 percent of the responsive documents. In fact, their searches had found only 20 percent.
The Superiority of Predictive Ranking
Here is an example from our case files that illustrates the savings from Predictive Ranking. A law firm received more than 1 million documents in a production from the other side. The lawyers ran keyword searches to find relevant documents. They found 200,000 documents, which they reviewed. Of those, 5,600 turned out to be relevant.
Using Insight Predict, we tested our Predictive Ranking process on those same documents, based on an initial seed set of sample documents. We ranked the 200,000 documents using the Predict algorithm to see how many documents the team would have had to review using our process.
The answer was 38,000 documents. The team would have found all 5,600 relevant documents in the first 38,000 they reviewed. They could have ignored more than 160,000 documents.
You can see the yield curve here:
The graph shows the number of documents that needed to be reviewed using the Predict process (red line) and linear review (gray line). At $2 per document for review (which is probably conservative), the savings would come to $320,000.
Savings in Time and Money
By using Predictive Ranking before starting your review, you can significantly reduce the number of documents to review, the time it will take to complete the review, and therefore the overall cost of the review. At the same time, you can complete the review with greater confidence that you have identified virtually all responsive documents.