Reducing E-Discovery Review Costs with Computerized Search
by Conrad Jacoby, Esq.
The cost of reviewing discovery document collections for relevance and privilege dwarfs just about every other segment of the litigation fact discovery process. Taking depositions? They consume almost no time compared to the effort required to review the average document collection. Expensive expert witnesses? Except in highly specialized situations, expert work, too, amounts to only a fraction of the cost of having law firm associates and paralegals review 100,000 (or 200,000 or 2 million) e-mail messages and loose documents for content.
Review costs are directly related to the number of documents that must be reviewed, so one of the most fundamental strategies for containing discovery costs is limiting the number of documents that must be reviewed. Search technology can help manage electronic document collection volume in several distinct ways.
I. Search To Remove Clearly Irrelevant Documents From The Review
When working with electronically stored information (“ESI”), digital materials that have been collected in discovery are often filtered to remove binary and other program and system files before any review begins. Filtering is somewhat the converse of search—it excludes documents rather than specifically identifying them—but it still uses search-like criteria and search technology to identify materials that do not require review. It’s an extremely effective way of removing significant amounts of irrelevant ESI from a review population. No review project should be started before this essential task has taken place.
However, even with the removal of executable and system files, it’s likely that a large amount of the remaining user-created files still have little if any relevance to the legal dispute. Human-directed search can help trim this population, too. For example, even at the start of a legal dispute, most legal teams have ideas about potential witnesses and potential phrases and vocabulary that might have been used in relevant e-mail messages and other documents. Running a search for documents that do not contain any of these terms yields a collection of documents that are less likely to be relevant in the case. Sampling or profiling these documents may yield analytics that can be used to exclude additional classes of documents from review.
For example, within collections of e-mail messages, it may be possible to carve out press releases, news digests, or other publicly available information from the review population because these messages all originate from a common e-mail address or Internet domain. It may also be possible to exclude common forms, such as invoices or automatically-generated reports that will not contain relevant information. The facts driving the legal dispute will determine what types of documents can be treated as an aggregate class using these techniques, but virtually every document collection will include documents that can be found—and excluded from review—using this search strategy.
II. Search To Divide A Review Document Population Into Meaningful Sub-Groups
Once documents collections have been slimmed down by filtering and targeted searching, computerized search can also be used to divide a review document collection into meaningful sub-groups. Searches can be used to identify documents that require urgent or heightened scrutiny, as in the case of privileged or “key player” documents. Finding and analyzing the content of these documents first often provides significant insight into the case, influencing case development—and settlement—strategy. If high priority documents are sufficiently compelling, a legal dispute may actually be resolved without requiring review of the full discovery document population.
Searches can also be used to divide review documents into specific categories to generally speed review. For example, reviewing e-mail messages in chronological order provides a scattered picture of activities and conversations that were taking place simultaneously. Using search to identify related documents reduces the time it takes for a review group to become fully knowledgeable in a discrete subject area, often greatly increasing speed of review and reducing the number of hours needed to complete that portion of the review. As an added bonus, even without using specialized duplicate and “near-duplicate” identification tools, search will naturally group closely related documents, permitting reviewers to quickly skim through documents that do not contain unique information.
Conclusion
Computerized search tools are a highly effective way to manage and reduce the cost of conducting any document review, whether large or small. Different cases will require a broad range of search strategies, some of which may only be useful inside a specific legal dispute. Time and time again, though, it’s been proven that a little bit of planning—and searching—makes a tremendous difference in how efficiently a review project can be run. For clients and law firms alike, that efficiency translates directly into lower cost for better review results.
About the Author
Conrad Jacoby is the founder of efficientEDD, a consultancy specializing in electronic discovery and litigation information management issues. A seasoned litigator as well as a technology consultant, Mr. Jacoby writes and speaks extensively on electronic discovery issues. He received his undergraduate degree, magna cum laude, from Yale University and received his law degree, cum laude, from the Georgetown University Law Center. He can be reached at conrad@efficientEDD.com.
Filed under Featured Articles, From the Experts, Home Page Featured, Technology Counsel.
November 20th, 2008 at 11:21 am
Key word searches are more tricky than one imagines. Big companies are international, using various languages (how do you say FCPA in French ?).
If you look for collusion, may some describe it as agreement, discussion, meeting.
Statisticians are only partly helpful; even though they can indicate relevance, they lack the domain expertise to pinpoint searches. BTW, lawyers themselves often lack domain expertise,since it is not a question of law to find relevant documents, but knowledge of business practices.