by Caroline Pollard - Ballard Spahr on January 11, 2017
Editor's Note: This article was originally published by Law Journal Newsletters.
Ballard Spahr is a national firm of more than 500 lawyers in 14 offices. We represent clients in more than 40 areas of litigation, business and finance, intellectual property, real estate, and public finance. Our clients include large public companies, privately held corporations, government agencies, nonprofit organizations, and the occasional individual.
Our e-discovery team uses technology to support clients in a cost-effective manner despite the exponential growth in data volumes and data types. A crucial part of our job is conducting early case assessment (ECA), a process that involves searching a collection and determining what data should be pushed on for review. Custom keywords and searches were the backbone of our ECA workflow. When we ran into a case where that workflow wasn’t accomplishing all it was meant to, we turned to a dedicated ECA tool and found it to be the right solution for our problem.
The ‘Overly Broad Collection’ Problem
One of our ongoing cases involves assisting a government agency in a contract dispute. The client provided us an overly broad collection of 2.8 million documents, which included full email accounts from multiple custodians and an overall low de-duplication rate. We were faced with a large amount of information that we knew was irrelevant to our case.
Although all of the documents had already been processed, we were in agreement with the case team that it was too much data. Our primary challenge was to find the most accurate and efficient workflow to cull the data, preserving all documents that would be important to the case.
In past cases, we would put irrelevant documents into a locked folder in a Relativity workspace. It was an imperfect system for us. These irrelevant documents would clutter server space and resurface during the review process. That meant that when our team pushed the case to review, we would have to remember to exclude those irrelevant documents when they cropped up again—just a further complication in an already complex process.
For this case, we began our ECA workflow in Relativity 9.1 using keyword searches, search terms reports, and pivot charts, hoping this would help us cull a good portion of irrelevant data. Considering how much material existed, however, that first layer of investigation was not in-depth enough for us to rule out a significant number of documents.
About this time, we decided to upgrade to Relativity 9.4. The upgrade enabled us to leverage Relativity ECA and Investigation. The shift gave us the tools we needed to improve our entire approach to ECA in this and future cases.
Unlocking New ECA and Investigation Abilities
Because analytics were always important in our ECA process, we were already on the right track going into the upgrade. Relativity ECA and Investigation allowed us to take work we had already completed and use it to customize our parameters going forward. We repurposed our initial list of search terms to cluster documents and built a visualization widget. Once we had that, we built a heat map of clusters that cross-referenced documents against domains and key custodian metadata.
Cluster visualization helped us more quickly and confidently identify small groupings of relevant documents. There were some clusters that appeared to be obviously irrelevant, such as clusters that contained Facebook, Twitter, and shopping references. Prior to our upgrade, we would probably have tagged the clusters as irrelevant. But, with cluster visualization, we were able to drill down through four layers of them. In the fourth layer, we discovered a small cluster of documents that appeared to be relevant. Without the cluster visualization option we might have “lost” the documents. Cluster visualization and the integration of widgets gave us much more insight into the data and helped us make more informed decisions.
In the end, we were able to find irrelevant documents more quickly, mark them up, and get them out of the way without accidently eliminating relevant data.
Faster ECA, User-Friendly Analysis
With this workflow, we’re more effective at culling and investigative workflows. When we prepare a case for review, we are confident that the data we push includes only the most relevant documents. The government contract dispute in particular demonstrated our ability to find important information more efficiently, allowing our entire team to act faster on the relevant issues. In that matter, we reduced a 2.8-million document case to 277,180 documents—less than 20 percent of the initial collection.
We now also have an even more impactful way to deliver our results to our case team and clients. Cluster visualization and heat mapping allow us to handle the data with incredible efficiency and generate visuals of the data that resonate with both our attorneys and clients. Visual data has become the most coherent way to share case progress.
Our case teams are happier too. The visual reports we now provide allow them to drill down into the data and lists of relevant documents, and we’ve noticed that they are more engaged with reporting than in the past when we simply sent them Excel reports. Even if we don’t have attorneys looking directly at Relativity, the customization option allows us to send them easy-to-read visuals instead of tables of information.
Where to Go with ECA and Investigation
Our team has three more ECA cases in motion. All three involve situations where the initial delivery from the client included entire email accounts. This almost always results in overly broad collection with a high rate of irrelevant documents. In one of these cases, a client used his email account for both personal and business communications. ECA allowed us to not only identify relevant business documents, but enabled us to give the client assurance that his personal data would remain personal.
We know the personal information is not relevant—a situation not dissimilar from the start of the government contract case. The difference between now and then is that we can use the analytics and visualizations to cull that data quickly and confidently from the jump.
At Ballard Spahr, we have a culture that revolves around self-sufficiency and a belief that technology allows us to assist our clients more efficiently. My team is at the forefront of that mission and, with Relativity as our e-discovery solution, we are in a position to continually build on the way we work and secure wins for our firm in the process.
Caroline Pollard, manager of e-discovery services at Ballard Spahr, has more than 25 years’ experience as a litigation support specialist. She currently oversees a team of four analysts and a media clerk in using e-discovery and technology tools to serve case teams and clients.