Three Drawbacks to Keyword Searches for Finding Electronic Evidence

The plethora of electronically stored information (ESI) being created on an annual basis creates challenges and problems for the legal profession which has to search for electronic evidence in a workplace investigation. In 2012, it is expected that 1.8 zetabytes of information will be created. This is the equivalent of every US citizen writing three tweets per minute for 26,976 years.

The legal profession, which is tasked with searching ESI, often uses outdated tools, which may not be as effective as newer technologies. Today, keyword searches are still the preferred method for locating ESI even though there are limitations and drawbacks.

A study conducted by Blair and Maron looked at the efficacy of keyword searches in locating discoverable information. The case involved a San Francisco Bay Area Rapid Transit accident. The discovery database contained 40,000 documents and 350,000 pages. While the attorneys believed that keyword searches located 75 per cent of relevant documents, Blair and Maron discovered that keyword searches located only 20 per cent of documents relevant to the litigation. Obviously, lawyers are at a significant disadvantage when the majority of documents relevant to an employment lawsuit are left undiscovered.

Synonyms, Descriptions and Misspellings

Keyword searches suffer from three main deficiencies:

1. Failure of imagination. It can be difficult to conjure up every conceivable synonym to describe a keyword in the litigation.
2. Guessing game. It can also be difficult to guess how the opposing party would have described the key terms in the case. The opposing party may have used acronyms or slang terms to describe key issues in the case.
3. Human error and misspellings. If a party is searching word documents and e-mails and has invented a perfect list of keywords, that perfect list may leave out common misspellings.

Technology, Statistics and Linguistics

Federal courts have caught on to the problems associated with keyword searches. For example, in United States v. O’Keefe, the magistrate judge found that “whether search terms of keywords will yield information sought is a complicated question involving the interplay, at least, of the science of computer technology, statistics, and linguistics” and is to “truly go where angels fear to tread.” United States v. O’Keefe, 537 F. Supp. 2d 14 (D.D.C. 2008).

The implication of the O’Keefe case is that a party using keyword searches to locate ESI should hire a special linguistics expert just to assist in the formulation of the searches. In Victor Stanley v. Creative Pipe, Inc., the magistrate judge found a waiver of the attorney-client privilege where the defendant failed to provide the court with meaningful information regarding the methodology used to select keywords. 250 F.R.D. 251 (D. Md. 2008).

Tips for Effective Keyword Searches

While the legal profession still relies almost exclusively on keyword searches, such a search methodology is not without its issues. If an attorney intends to rely on keyword searches to locate ESI, the following strategies should be considered:

• Have a cooperative relationship with opposing counsel to determine how the other side describes its search terms.
• Document properly the methodology used in formulating keyword searches so there is no waiver of the attorney-client privilege.
• Assign a person to review the remainder of the documents to determine if there are privileged documents left in the set. Given the problems associated with keyword searches, it is not wise to use keyword searches as a privilege review tool.
• Depending on the volume of ESI in the case, consider predictive coding as a viable tool in lieu of keyword searches where appropriate.

Look for more on predictive coding in my November blog post.

Originally published on i-Sight.com's blog: “3 Drawbacks to Keyword Searches for Finding Electronic Evidence.”