Law departments face enormous costs and risks in managing electronic discovery. Many reported decisions show what can go wrong in EDD. Many law departments have horror stories about out-of-control costs. It does not have to be this way. Relying on discovery experts minimizes both costs and risks. Barrasso Consulting, LLC, with 200+ litigation professionals, specializes in managing the document collection and review process. Diane Barrasso, the founder, explains in this interview with Ron Friedmann of Prism Legal Consulting, Inc., what makes her company's approach so unusual and valuable.
Friedmann: What does Barrasso Consulting do?
Barrasso: We manage the discovery process to reduce cost and risk.
Friedmann: How did you get into this business?
Barrasso: In the 1980s I was a microbiologist. A law firm retained me as an expert witness. I was comfortable with computers at a time when most lawyers found them mysterious. The lawyers were intrigued by my ability to computerize timelines and adjust them instantly. I did some more expert witness work and then moved into the discovery management business.
Friedmann: Discovery is costly. What's the best way to get a handle on it?
Barrasso: Gaining control of the review process is critical. The biggest cost is lawyer time to review documents. Putting less irrelevant information into the discovery pipeline can dramatically reduce costs. Smart approaches to review also include using appropriate tools to gain efficiencies in the process. Of course, discovery does not start with review. You start by collecting documents. The challenge there is to limit the scope of what you collect while maintaining defensibility at the same time. Using cost-effective technologies to further limit the scope, such as searching, de-duplication and other filtering, adds the final component of a solid discovery program.
Friedmann: Have you found ways to improve document review?
Barrasso: I've analyzed review in detail. My philosophy is that context-based approaches are best practices for document review. With a clear set of rules, training based on those rules, and monitoring throughout, a high-quality, low cost document review is possible.
Friedmann: What do you mean by context and rules?
Barrasso: We train reviewers on the client's industry, its organizational structure, the main regulations governing it, and how the company acts in the ordinary course of business. By understanding the context, reviewers have a much better sense of how a given document fits into the review equation of what is responsive and what is privileged. We also drill down to details. For example, we rigorously seek out project nicknames and also learn how the client's employees involve legal in their daily activities. We also learn to be aware of trade secrets and other sensitive documents. Context is not enough though. Without clear rules, even smart and experienced lawyers treat documents differently. We write detailed rules on how to review documents. Training on these rules takes time upfront but speeds the review rate and improves consistency.
Friedmann: You also mentioned the overall process. How do you manage that?
Barrasso: A clearly defined workflow, project management, and metrics are essential to successful reviews. For example, we measure documents reviewed per hour by reviewer. That lets us evaluate each reviewer and plan for the resources we need to finish on time. We look at parameters such as review rates, error rates, redaction requirements, and volumes of trade secrets. We also quantify differences among matter types, for example, patent, investigations, employment, commercial litigation, and securities. The variability is surprisingly high. Cases with higher levels of complexity mean slower reviews and we plan budgets, scopes of collection, and staffing accordingly.
Friedmann: How does technology factor into what you do?
Barrasso: We're not a technology company so we are vendor neutral. But with the constant stream of new technology, we regularly evaluate new tools. Our analysis is rigorous, with both quantitative and qualitative methods. We've concluded that there is no single perfect software for all cases. Our extensive hands-on experience with both market-leading and niche products means we can help clients determine what's best for a case. I believe we have a unique perspective.
Friedmann: So you always use outside technology?
Barrasso: No, we have created our own technology to manage the review process workflow. It collects metrics and generates statistical reports. It works with any document review tool. We use it to distribute work to reviewers and monitor productivity. Building it was a big investment but it has paid off in added efficiency and quality. We also have technology for managing the audit trail of a document collection effort. Both tools have been refined over many years of rigorous use by our team.
Friedmann: Do you ever use two or more tools on one case, say concept searching plus email analytics?
Barrasso: Usually we use just one, but a single tool may offer many approaches within the tool. Email analytics is great for early case assessment and for tracking conversation threads. But for hard core document review, we find that the document file path is the best information available - it's the most valuable in reducing the size of a collection significantly. (The file path shows where a computer file is stored, for example: M:\Shared\Marketing\Advertising\Campaign\Slogans\draftslogans2.doc.)
Friedmann: Wow, the file path was in DOS. I'm surprised to hear it's so important.
Barrasso: Big companies limit how much email users can keep in their mailboxes. So users frequently move email from Microsoft Exchange servers to their hard drives. In doing so, most users organize their messages into file folders according to topic or concept. And all the information in the path provides a lot of context about the file, especially when combined with the document title or message subject line.
Friedmann: Will technology replace human review in the future?
Barrasso: That's unlikely to happen any time soon. Artificial intelligence has limits. First, we still don't know the complete judicial reaction. And second, we've found that getting the most out of concept searching or other artificial intelligence is costly. For each case, you have to "teach" software all the guidelines or hire an army of linguists. Then you have to test with samples. With all that prep, if you just rely on people, you can be done with the document review quickly and usually at lower cost. Advanced tools help but I don't foresee a single solution - the right software will always depend on case needs.
Friedmann: It sounds like you have a great handle on review. You also said the collection process is important. What's your approach there?
Barrasso: Many lawyers have over-reacted to widely publicized discovery sanctions. There's a tendency to "back-up the truck to the loading dock and collect everything." This is counter-productive because it does not distinguish between collecting and preserving information. When we're involved from the onset of a case, we help clients understand that distinction and minimize the volume collected.
Friedmann: How do you minimize the volume?
Barrasso: Sometimes we use key words, but this can yield too much data and usually requires adversary approval. The best approach is talking to employees to understand how they have filed documents. Employees typically have their own filing systems. Talking to them, we can determine what's stored where. Knowing this dramatically reduces the volume and increases the relevance.
Friedmann: Is this reliable and legally defensible?
Barrasso: We start by talking to IT to understand the range of places where employees can store data. But IT doesn't know how individuals store their own data. When we meet with custodians, we follow a script. We have a conversation about their data. The defensibility comes in by having a consistent, well-documented, good-faith approach that is reproducible.
Friedmann: Can you put numbers on the reduction in volume you achieve?
Barrasso: Statistics drive much of what we do. Our relevancy rates are often 75% to 85% for cases where we do the collection. Where we're called in after the document collection to help with the review, we find relevancy rates as low as 10%. This roughly eight-fold reduction does not reduce the number of responsive or produced documents. When you collect every drive, every server, every memory stick, there's a lot of irrelevant information. Talking to custodians first to understand what they've stored and where simply makes sense.
Friedmann: Is there a "secret sauce" for a significant reduction?
Barrasso: We employ experienced project managers who customize the collection process for each case on our backbone of template tools. We also select our attorneys for their communication skills and train them extensively. They use a template for collection, modified for each case. That provides a structure for a careful and investigative yet friendly dialogue with custodians.
Friedmann: How do law firms react to your approach?
Barrasso: Some law firms don't like to lose control, but once we convince them we make discovery supervision easy and take the worrying out of the process, they prefer to use us for document work. We make special efforts to partner with and involve firms. I should add that we don't have the same stake in the case as law firms. We set objective guidelines, which we follow closely. This is very attractive to General Counsel.
Friedmann: We've covered a lot. What's a key point that readers should take away?
Barrasso: Law practice has many sub-specialties. For example, within litigation you have antitrust, patents, real estate, and appellate work among others. Clients understand the need for specialists in law practice. Just because a law firm or department has substantive legal expertise does not mean they have the same process expertise in document discovery. Clients should recognize that document discovery is its own specialty, and therefore it can be effectively uncoupled from other litigation activities.
Published September 1, 2007.
