Keysearch Clarity

There is nothing more sacred to in-house counsel than attorney-client privilege. But with the advent of e-discovery, preserving privilege has become increasingly difficult. A recent court ruling in the District of Maryland underscores this point.

On May 29, Judge Paul Grimm, noted as one of the most influential and educated jurists on the topic of e-discovery, handed down an opinion that many experts consider the most significant guidance to date on preserving privilege in the course of e-discovery.

In Victor Stanley Inc. v. Creative Pipe Inc., the plaintiff filed suit alleging intellectual property infringement and unfair competition, claiming the company lost contracts because the defendant inappropriately accessed images from its Web site

Victor Stanley made an e-discovery document request. As is mandated by the amended Federal Rules of Civil Procedure, the two parties sat down at a "meet and confer" conference to outline the specifics of the request and the production. Although they came to an agreement regarding the discovery's methodology, the defendant abandoned it after the court extended the production deadline by four months. Instead, the defendant--without consulting the court or the plaintiff--ran a keyword search to identify privileged documents. They inadvertently produced 165 privileged documents.

Despite the defendants' efforts to convince the court to compel the opposing party to return these documents, Judge Grimm ruled that because the defendants failed to take reasonable precautions to protect privilege, they voluntarily waived privilege. "Judge Grimm's opinion is the first guidance we have for nonhuman privilege review," says Laurie Weiss, a partner at Fulbright & Jaworski. "Because keysearch terms are becoming more critical in evaluating large volumes of data, this case is especially significant."


Keyword Crisis
A series of mistakes, shortsightedness and ignorance sank Creative Pipe's argument that Victor Stanley should return the 165 privileged documents. First, during their meeting, Creative Pipe requested a clawback agreement from the court. A clawback agreement is an agreement between the two parties involved in e-discovery providing that inadvertently produced privileged information may not be used. However, the defendants, believing the four-month extension would make a clawback unnecessary, abandoned this request.

Second, after the extension, Creative Pipe told the plaintiffs and the court that the extra time would allow it to embark on a human-led privilege review. This means that it was going to use human reviewers to read through the corpus of potentially responsive documents to identify privileged materials. However the defendant abandoned this tactic and decided to use a keysearch methodology instead.

When the inadvertent disclosure came to light, Creative Pipe argued the documents should be returned. Judge Grimm put the spotlight on the defendants to explain their search methodology to prove a certain level of reasonableness and care was taken. "The defense was vague in its response," says Patrick Burke, associate general counsel at Guidance Software Inc., a provider of forensic data acquisition and analysis software. "They should have had a process in place, followed that process and then documented it. Rather, they couldn't explain their search methodology."

Testing and Sampling
All too often producing parties lack an understanding of what comprises their search methodologies. Judge Grimm's ruling in this case shows that ignorance is not a reasonable excuse.

"The vast majority of folks are operating with a black box search," says Dean Gonsowski, vice president of e-discovery services at Clearwell Systems Inc., a provider of e-discovery software.

Victor Stanley is a lesson to counsel that it's not just about successfully finding privileged documents via keyword searches; it's about showing your work. And the most commonly accepted method is through testing and sampling.

The process begins with the development of a list of keysearch terms. These terms should be run against the corpus of data. Human reviewers should then look at a random sample of the data to see the effectiveness of the keysearch terms. "If you have 100,000 documents, you should look at 500 to see what kinds of results you're getting," Burke says. "You might say, 'This test criteria missed this when we used these keywords,' in which case you can then go back and adjust your keysearch terms to get more effective results."

The second part of this process is sampling the body of data that did not come up in the search to ensure the search software isn't passing up any potentially responsive information.

"What you know is what the search terms have caught; what you don't know is what they haven't," Weiss says. "In the context of privileged documents, that is the arena where litigants can get in trouble."

Counsel should sample the documents not picked up by the search, and review them for responsiveness and privilege to ensure that the search terms are returning meaningful results. If not, then lawyers should tweak the search terms to maximize their effectiveness.


Ripple Effect

Although the opinion in Victor Stanley is technically confined to a small jurisdiction, experts think Judge Grimm's ruling will have dramatic implications on a national level.

"Decisions like this from a tech-savvy judge like Grimm can be relied upon by other judges who are less tech savvy," Burke says. "When those less tech-savvy judges read something as clear as this ruling, they can make more confident choices about what is reasonable and what isn't."

In the end, this type of guidance will likely help in-house counsel prevent their companies from becoming the next example in court. "The more decisions like this that provide guidance on what is a reasonable e-discovery process, the more confident in-house and outside counsel can be in setting up a procedure that will be reasonable," Burke says.

Technology Editor

Keith Ecker

Bio and more articles

Join the Conversation

Advertisement. Closing in 15 seconds.