Mining Mountains of Evidence

Though dramatic courtroom scenes in TV shows and movies might suggest otherwise, legal cases often hinge less on surprise witnesses, and more on sifting through towering stacks of documents. In the past, this process was not only slow, but given that the people reviewing the documents were highly-paid attorneys, extremely expensive.

Flash-forward to today and the situation is changing. Law firms are taming the torrent of information by using predictive analytics solutions such as IBM® SPSS® Modeler to automatically cull through millions of documents in minutes, tag them for potential relevancy and duplication, and help attorneys complete their reviews in a fraction of the time.

“With IBM SPSS Predictive Analytics, we’re really cutting a lot of cost, increasing accuracy, and reducing the time it takes to do a review,” says Gerard J. Britton, director of investigative and compliance services at CAAS LLC, a company based in New York and Washington D.C. that provides technology solutions and consulting services to law firms, law enforcement agencies and other clients. “It’s really a home run application.”

Changing the economics of e-discovery
Lawyers have long dealt with evidence “discovery” – the early phase of litigation when each side of a dispute supplies the other with pertinent evidence and documentation, usually paper records. The advent of electronic documents, however, has spawned the new field of “e-discovery” in which lawyers comb through millions of emails, PDFs, spreadsheets and other computer-based information to find relevant evidence. And as the volume of electronic documents exploded in recent years, e-discovery became so labor intensive that costs skyrocketed – sometimes running into hundreds of thousands of dollars for a major case.

“E-discovery has created a cost component within litigation that in some cases can exceed 50 percent of the total cost of litigation,” says Britton. A big part of the problem was that the pile often contained hundreds or even thousands of duplicate or near-duplicate documents, but in no particular order. Wading through this jumble was slow going, with the average attorney able to scan only about 50 documents an hour.

Enter IBM SPSS Modeler and its built-in text analytics capability, which is boosting the speed and efficiency of e-discovery by instantly sorting the piles of evidence into orderly groups of virtually identical documents. “If you have a million documents, there may be clusters of thousands of almost identical documents,” Britton says. “Using the IBM SPSS solution, you can identify and pull together those documents that meet some threshold of being nearly identical.”

As a result, Britton says, “One person can spend 20 minutes on the first document, and quickly review the remaining variations to see there’s nothing that changes their nature.” What’s more, attorneys actually like having the stacks pre-sorted in this manner. “Review attorneys actually find it more relaxing,” he says. “They’re getting sets of documents that they can just scan and find what they need, instead of dealing with a cacophony of documents coming from every direction.”

Significant savings, greater accuracy
Britton estimates that IBM SPSS Modeler has enabled attorneys to jump from reviewing about 50 documents an hour to 120 or more, depending on the nature of the document set. The efficiency boost is generating significant savings, usually reducing e-discovery costs by 30 to 70 percent. Depending on the case, Britton says, “a $100,000 bill for e-discovery may be reduced to a $60,000 bill, and a $500,000 bill reduced to between $200,000 and $300,000.”

CAAS’s clients appreciate the savings. Indeed, in the current tough economy, they’ve come to expect nothing less. “Clients will not tolerate paying more than they have to for discovery,” Britton says. The solution is also breathing new life into cases that may have been non-startersdue to high discovery costs. “It’s often said in the legal industry that the decision whether or not to litigate turns not on the merits of the case, but on the relative cost of your discovery,” he explains.

Not only is Modeler helping CAAS speed the document review process, but it’s also boosting the overall accuracy of the documents identified as relevant. The key is a better quality control process in which CAAS employs Modeler’s text-coding tools to predict which documents are likely to be germane to the law suit. The firm’s review managers then compare this prediction against the set of documents actually selected by the reviewing attorneys. By zeroing in on those instances where the model-generated document code varies from the reviewers coding decision, the firm can perform quality checks that are significantly more effective than reviewing a random sample of documents.

Transactional analysis
Britton worked with IBM SPSS Predictive Analytics solutions even before he arrived at the firm. As a prosecutor, he used Modeler and IBM® SPSS® Statistics to manage and analyze large volumes of transactional data. For example, in the case of a company that is preparing itself to be sold, Britton says, there might be a suspicion of “channel-stuffing” – when companies intentionally inflate sales figures – before the announcement of quarterly earnings. Using predictive analytics, firms can quickly scan thousands of financial transactions and spot the earmarks of this behavior. And by doing this number-crunching at the beginning of case, law firms and their clients can develop an early view of what lies ahead.

The results help CAAS decide how to proceed with the investigation, or whether to proceed at all. If the refined data supports the client’s claim, Britton says, the investigation can proceed apace. This early legwork also ensures that when a hired expert enters the picture, he or she can proceed with much sharper – and therefore more economical – focus.

Complete solution
CAAS’s e-discovery solution integrates IBM SPSS Predictive Analytics with industry-standard review tools and best practices, creating a complete solution that law firms of any size can afford. “IBM SPSS Modeler is a tool that can be inexpensively owned by a stakeholder,” Britton says. “It’s not overly cumbersome, and you can create very elegant and very inexpensive solutions that cut down on cost.” The solution is also flexible, allowing law firms to fine tune the system by incorporating the specialized knowledge
of “domain experts” to make the solution more effective in the real world.

CAAS is now helping firms bring this comprehensive predictive analytics solution in-house. “Using IBM SPSS Modeler, we’ve created this solution that enables firms to be self reliant,” Britton says. “They won’t have to rely on what an e-discovery vendor charges. They will also have a toolkit, which will lend itself of all kinds of creative solutions. This is a beautiful thing in an idiosyncratic environment like litigation.”

About IBM Business Analytics
IBM Business Analytics software delivers complete, consistent and accurate information that decision-makers trust to improve business performance. A comprehensive portfolio of business intelligence, advanced analytics, financial performance and strategy
management and analytic applications gives you clear, immediate and actionable insights into current performance and the ability to predict future outcomes.

Combined with rich industry solutions, proven practices and professional services, organizations of every size can drive the highest IT productivity and deliver better results.

For more information
For further information or to reach a representative: www.ibm.com/software/analytics/spss

Request a call
To request a call or to ask a question, go to www.ibm.com/spss/contactus. An IBM representative will respond to your inquiry.