Containing Costs with Concept Searching Saves Clients Thousands

As my firm's Technology Partner, I oversee the e-discovery process for most of the large cases at Severson & Werson. In 2008, one of our clients, a major national bank, was sued by its recently acquired mortgage lending company just after the subprime mortgage financing collapse. The case hinged upon the interpretation of a term in a contract related to the valuation of certain types of subprime loans in the securitization process.

We identified a set of key documents and concepts in our possession, which we had to use to find similar documents in the opposing party's production. We faced, however, multiple challenges in trying to do so.

One challenge was time. We were under pressure to find these documents quickly in order to conduct depositions and stay on track with our litigation schedule.

To make matters worse, despite agreements made regarding e-discovery during the Rule 26(f) conference, the opposing party failed to cull down their data set. Instead, they produced 80,000 e-mails, many of which were irrelevant to the litigation.

Finally, most of these documents lacked critical e-mail metadata fields we requested, making it exceedingly difficult to match our key documents to similar documents in the lender's production or to track them.

Search for a Solution

We tried to compel the lender to produce the metadata. But the process of compelling the opposing party was met with lengthy debate and hearings with the discovery referee. With depositions looming, we could not remain idle while waiting for the motion to be granted and risk falling behind our litigation schedule. Therefore, we had to find an alternative. Specifically, we wanted a solution that would allow us to identify these key documents within the massive data set without skyrocketing legal costs for our client or a large investment of time.

Using a service provider, we initially attempted to minimize the document collection and pull together documents by running a near-duplication detection. This failed, however, for several reasons. First, despite the fact that we had clear examples of the documents and language for which we were looking, near duplication did not give us the ability to run a targeted search for just the documents or clauses we were seeking. Instead, it was a much more general approach that did not leverage our knowledge of the issues.

Text Analytics Assistance

Frustrated, we contacted Evolve Discovery, an e-discovery consultancy, to help us find a solution that could match key documents from our production and the lender's. We also needed an efficient means of reviewing 80,000 e-mails. In addition, we needed the solution to be cost-effective, efficient and scalable, as we had multiple offices and multiple parties that required access to the case information.

The consultancy recommended using text analytics to narrow the number of documents to a more manageable set. We decided to go with Relativity, a Web-based litigation support solution developed by kCura Corporation that has a built-in text analytics package that includes the ability to identify similar documents based on concepts and to perform concept searches. Evolve Discovery hosted the solution.

We decided not to purchase the software and to go with a hosting solution because we did not want to make a large investment of our own resources. With the Web-based model, no internal resources were required of us other than the use of our PCs. There was also no business interruption for our IT staff as the solution required no assistance on their part. An additional benefit happened to be our hosting provider's expertise as an e-discovery consultant. During the litigation, one of its professionals spoke before a special discovery master as an expert, testifying with regard to the importance of metadata production.

Like most attorneys and paralegals, our legal staff did not have the time or the desire to learn a new piece of software. Yet to harness the full potential of the solution, training was necessary. Fortunately, getting our personnel up-to-speed on Relativity was completed within a day: 45 minutes to be exact. A small learning curve was critical in saving us time while containing costs for our client.

Using traditional culling methods had already fallen short of our goal of finding key documents. The solution came equipped with embedded text analytics capabilities. The text analytics allowed us to search the data set based on concepts, thus leveraging our knowledge of the facts and making use of the key documents we already found. All we had to do was select relevant lines of text in our key documents ' specifically clauses regarding valuation of subprime loans ' right-click and run a concept search against the opponent's production. The software automatically returned documents that contained similar concepts, even if the wording was not identical or there were variations in the content. In the end, we found about 1,000 documents similar to our original examples, so we focused our efforts on this data.

Easy Review and Scalability

Although we quickly honed in on the most important documents in the case, we still were tasked with reviewing the entirety of the lender's production. Working with a $100,000 budget, we had 10 review attorneys and little time.

Fortunately, Relativity's user interface was much more intuitive than other applications we had used in the past. Unlike previous applications, all we needed to do was click a tab to automatically display all documents that had not yet been reviewed. The solution also supported load files from all major litigation support applications, including Concordance, which we had previously used. This allowed us to transfer all work product from Concordance and integrate it seamlessly. We also were able to track all of the review attorneys, and we performed quality checks on all the reviewers as they progressed through the documents. In all, the review took only 10 days, and we ended up using half of our budget, saving our client $50,000 in review costs alone.

In addition, the solution's scalability enabled a diverse team that included attorneys in San Francisco and Irvine, CA, to access the same materials and collaborate all within the system. In fact, we also had experts, vendors, witnesses and representatives from our client accessing the system, totaling around 40 people. There were no issues of security because the software enabled us to set permissions so certain information was only viewable to certain people.

Favorable Outcomes and Cost Containment

In the end, after we found the critical documents and reviewed the large set of e-mails provided to us by the lender, we were able to step back and gauge the weaknesses and strengths of our client's case. Our case proved to be strong, and so we approached the opposition to negotiate a settlement.

Without the benefit of the text analytics features within the software, we never would have gained early insight into the case. Without this insight, both parties would have had to produce information from an additional 33 custodians, nearly twice as many as the number both parties had already produced. Incorporating more custodians, in turn, would have necessitated the review of thousands more documents. By taking advantage of concept search technology and choosing a product that was intuitive and scalable, we earned our client a favorable settlement and saved them potentially hundreds of thousands of dollars.

Key Lessons Learned

Choose a tool with scalability. This case necessitated that we collaborate with various parties, from witnesses, to experts, to our client. To work efficiently, we needed a tool that would allow us to grant simultaneous, secure access to the data. The data also needed to be accessible via the Web, with the necessary ethical walls in place.