Even companies that truly want to fulfill their e-discovery obligations properly are struggling to figure out exactly how to get the job done. Indeed, in recent months, the topic of e-discovery "search" has taken center stage. Traditional approaches to sifting through large collections of data for relevant or privileged information, such as keyword and Boolean search, are being called into question by some members of the bar, the bench, and the industry at large.
Seemingly overnight, the once familiar practice of document review in discovery has emerged as a new expert discipline. Courts rightfully are concluding that search in e-discovery can be a complex, scientific exercise. E-discovery in some cases may even require expertise in fields such as linguistics, statistics, and computer science.
In one of the most important e-discovery cases this year, Victor Stanley, Inc. v. Creative Pipe, Inc., Magistrate Judge Paul Grimm of the District Court of Maryland not only characterized search as an "information retrieval" problem, but also raised the bar for lawyers and litigants, requiring them to prove (e.g., through quality assurance testing and measurement protocols) that the search methods they employ are reasonable.
Meanwhile, dozens of vendors and consultants purport to offer cure-all search technology, but their claims are not necessarily backed by proof. How are litigants supposed to defend their search methods if their technology vendors and service providers cannot do so themselves?
Well, new developments in industry and government are taking shape to address this need.
For example, the National Institute of Standards and Technology and the U.S. Department of Defense created the Text Retrieval Conference, or TREC, as far back as 1992 to support research on text retrieval methodologies. TREC formed its Legal Track more recently in 2006 to meet the exploding demands of e-discovery.
The Legal Track seeks "to apply objective benchmark criteria for comparing search technologies" in the context of e-discovery, and to discuss capabilities and limitations of various automated approaches to e-discovery. The expected result is a concrete reference by which to evaluate competing technologies and vendors.
The profile of the Legal Track rose considerably this year with the introduction of a new testing protocol patterned after the real-world demands of litigants in discovery. The new protocol was designed to make the results more practical and to attract broader participation. Eighteen participants have responded by entering the study.
H5, a specialized litigation search and retrieval outfit based in San Francisco, is an example of one of the participants in this year's Legal Track. H5 has been a long-time advocate of TREC, having espoused and incorporated its protocols well before the Legal Track was even founded.
H5 views its participation as an opportunity to be at the forefront of a standard-setting movement - one with potential to become an authoritative protocol for document search and review in legal discovery.
It is not yet clear whether TREC actually will emerge as the definitive standard for testing e-discovery claims. With the support of government, academia and industry alike, it certainly stands a fair chance, and if nothing else, its existence is a positive step in the right direction.
Indeed, at the very least, the initiative hopefully will encourage companies engaged in e-discovery to put vendors to the test - by questioning their claims, demanding explanations as to how their processes and technology work, and requiring objective proof of performance.