In the year 2000—the early days of e-discovery—a case with 300 gigabytes of data was considered huge. Today? That’s just another standard e-discovery project. And while it’s true that multi-terabyte- and petabyte-sized matters can be a strain because of their size, most of today’s real e-discovery challenges come from the variety of data included in a case. With social media, text messages, video, audio, and financial transactional data now considered standard in many investigations, counsel must consider how to uncover a single incriminating text message from hundreds of different mobile devices for high-stakes litigation, or conduct a full 360-degree reconstruction, including chat, audio, and transaction records, for a regulatory investigation into a financial services institution.
Skip Walter, Chief Product Officer for FTI Consulting’s Ringtail e-discovery software sat down with me recently and discussed how visual analytics and e-discovery have evolved over the last 20 years. Skip was a pioneer in the use of visual analytics in e-discovery. During our conversation, I asked him about the excitement of those early days, how visual analytics are helping solve today’s new challenges, and what the next generation of e-discovery software could look like.
JR Jenkins (JR): In 2000, banker’s boxes full of paper were still the dominant “document type,” with electronic documents and email only just starting to become standard business tools. How did you make the leap from paper to “visual analytics?”
Skip Walter (SW): I started working with visual analytics in the late 1960s. And from the start, they showed real promise for improving individual decision making and problem-solving in a wide variety of businesses. Typically, people rely on their left brain for problem-solving, but that side of the brain is much slower than the right side, which is adept at recognizing patterns. I wanted to find a way for people to tap into that “right-brain” ability. About every five years, I would develop a new prototype to apply interactive visualizations to large bodies of text, but the “right” use case remained elusive.
By the early 2000s, I had become very interested in the challenge of making sense of unstructured (mostly textual) data. Early players in visual analytics were focused almost exclusively on structured data (the data residing in a database), which left a huge knowledge gap inside the corporation as the majority of the “intellectual property” and “business intelligence” is locked away in email servers, file shares, and Microsoft Office documents.
JR: Thus, the start-ups of the early 2000s focused on business intelligence, knowledge management, content management, etc.?
SW: Exactly. But while BI could be found by analyzing transactional databases, which helped companies improve supply chain, inventory, pricing, etc., ROI was always hard to find in the unstructured data if you were just trying to find “knowledge” inside those systems. The “aha” moment came when I was introduced to the problems of electronic discovery by some attorneys from K&L Gates. After just a few hours looking at the problem space, it occurred me to me: legal review was the perfect use case for visual analytics.
JR: This is a time when keyword search—and search terms—were really the best (and perhaps, only) method to find, organize, and manage an electronic document review. Wasn’t that enough?
SW: Not really. The most challenging problem to solve in e-discovery was finding the 2% of the relevant documents inside the 98% not relevant to the case. And with Boolean search, which has long been the de facto method for finding information, you have to know what you are trying to find at the outset. What made visual analytics so appealing is that it would allow attorneys to find relevant information with very little prior knowledge of the matter and the case details. In the early days of e-discovery, this was a completely novel way of sorting out the content of a case and getting to the responsive documents quickly.
JR: But this didn’t set off a race to develop the best visual analytics in e-discovery, did it?
SW: No. Much to our surprise, there hasn’t been as much competition in the visual analytics space as I expected. The types of visualizations that are used haven’t changed much over the past 20 years, and in many ways, the technology has not broken out of the legal field to use in other areas of business. Boolean search is still often used in e-discovery scenarios that would benefit from visual analytics. It can be difficult for lawyers to get over the hurdle of how to use visual analytics and appreciate the power visualizations can deliver. The use of visualization technology does not come naturally to most lawyers—their strength is typically with words—and that continues to be a barrier to adoption.
JR: So, what did change?
SW: What has changed is the technology’s ability to visualize new types of data in some amazing ways. Essentially, we have expanded the scope of what we can visualize well beyond views into “structured” data and “unstructured data” and are starting to see integrated blends of the two. One great example is how counsel can now build a 360-degree reconstruction of a series of financial transactions overlaid with social network visualizations for a fraud or regulatory investigation. Visual analytics can capture and display in multiple ways the connections between phone logs, chat data, financial transactions, and other forms of time-stamped communication between key persons of interest. The software has the unique capability to stay apace with the evolving data landscape.
JR: To your point about adoption barriers, what are some of the ways you educated lawyers about visual analytics in the early days? How do those differ from today?
SW: Most people don’t remember this, but it took about five years before spreadsheets were widely adopted for business use. A screenshot of a spreadsheet does nothing to show the user all the things the software can do. But with a demo, the user quickly recognizes the many ways it can help make their jobs easier. It was the same with visual analytics. As soon as a lawyer was able to see a live demo comparing and contrasting between linear review and visual analytic review, they began to understand how much more productive and cost-effective their e-discovery projects could be.
That is still where we are today. There are so many opportunities—keyword validation, trial preparation, investigations, fraud detection, FOIA requests, compliance, information governance, etc.—for visual analytics to save time and money. The legal industry is becoming increasingly receptive to the possible benefits. But at the end of the day, the review managers and e-discovery leaders need to see it in action to believe the business difference it can make.
JR: What are your predictions for how visual analytics will impact the legal world in the foreseeable future?
SW: Efficiencies and capabilities we now take for granted in e-discovery would be extremely useful in other areas of business. There are a lot of business ecosystems, including Microsoft Office 365 and Salesforce, that can make use of visual analytics. I think we’ll start to see those applications taking shape in the near future. The first change we will see is the expansion of visual analytics from text to video, audio and images and then into structured data. I’m also very excited about the next generation of hardware and opportunities to incorporate visual analytics software with augmented reality, virtual reality, and mixed reality scenarios. VR hardware is beginning to expand beyond video gaming capabilities, and will soon give us the ability to visualize patterns across hundreds of millions of documents. We will be able to see patterns in a corporation’s entire litigation portfolio at one time, or all the enforcement actions across a regulatory agency’s history. The scale is stunningly large, and the possibilities for how this information may be used to improve processes are limited only by our imagination.