Dave Walton is among the rarest of lawyers, one who is dedicated to expanding the use of big data and predictive analytics in the legal field. It’s a tough challenge because the sector is inherently risk-averse. But Walton, chair of Cyber Solutions & Data Strategies at Philadelphia-based Cozen O’Connor, is determined to effect change in the sector.
Wharton marketing professor Raghuram Iyengar, who is faculty director of Wharton Customer Analytics, recently joined Walton for an interview with Knowledge at Wharton about the challenges of getting lawyers to embrace analytics, and the benefits they stand to gain. Iyengar also teaches an executive education program titled, Customer Analytics for Growth Using Machine Learning, AI, and Big Data.
An edited transcript of the conversation follows.
Knowledge at Wharton: The legal field layers on top of just about anything we can imagine, which creates a lot of analytics challenges. What are some of those challenges?
Dave Walton: I think the legal field is still in its infancy on analytics and big data…. We’re still trying to figure it out, and there’s still a lot of consternation in some corners about what is analytics? What does it mean to be a lawyer? Lawyers have this idea that, “Well, my brain was trained. My judgment is everything. My personal experience is everything. There is no way a computer could ever do my job.” That’s what they mistake analytics and AI for. A lot of lawyers don’t understand it’s using data to supplement your judgment, your experience and your decision-making process, and perhaps seeing things that you wouldn’t otherwise see because you have access to data analytics and data expertise.
Knowledge at Wharton: I imagine there are both client- and non-client-facing impacts because lawyers have to think about how it could enhance their work, but they also have clients who are creating these immense data footprints that weren’t there even five years ago.
Walton: Absolutely. Clients are ahead of us in using data, and now they are looking at law firms to say, “How are you going to use your data to better provide services to me?” But there’s an internal tension there because a lot of law firms are still billing in the billable hour. There isn’t necessarily an incentive to be efficient economically. Clients are now pushing that incentive to us, making us do alternative fee agreements where the use of analytics can really help internally.
Externally, you’re starting to see some movement on how to use publicly available data in combination with client data and in combination with law firm data on more of the predictive analytics. Can you predict the outcome of a lawsuit based upon data? I think a lot of people would say, “You’re crazy for even thinking that’s possible.” But I think there are opportunities to use all different types of data sources, most of which are public, in a way that enhances a lawyer’s ability to predict not only the outcome of the case, but what it’s going to cost and how long it’s going to take to resolve.
Knowledge at Wharton: Raghu, what do you think about these issues?
“Clients are ahead of us in using data, and now they are looking at law firms to say, ‘How are you going to use your data to better provide services to me?’” –Dave Walton
Raghuram Iyengar: A lot of the things about legal analytics, I actually learned from Dave’s writings. What he brings up in his writings is that, regardless if you’re a law firm or a firm that is working with big data, you should think carefully about the implications in terms of hiring, for example. What data are you using for customer hiring? What data are you using for targeting customers? Where are the data residing? Who has access to that data?
I think these are key questions. Whether you are a legal firm or whether you’re using analytics for many different purposes, you have to answer them. The question for many of these companies is, what are the legalities around it? As Dave mentioned, many of these things are in flux. For instance, California recently passed a law that basically said there are certain things that companies residing in California could do with their customer data. A big question for these companies is, are the customers residing in California? What should their data be? Should it be present only in California? All of this is in flux.
Knowledge at Wharton: Access to data has made us more global than we’ve ever been, yet there are no global regulations on data. As a result, different regulations are often in conflict with one another. What can we do about that?
Walton: That shows the crossover in the different spaces that we’re talking about here, because you’re talking about data crossing over into privacy law, into cyber security law. It’s always going to be balkanized, so I don’t think you’re going to have a global data law. Even [the E.U.’s] General Data Protection Regulation is not a monolithic law. Every participating state still has to adopt their version of what the GDPR regs are. There are differences, and it’s hard to navigate. Then you get into Africa, and you get into Asia, you get into other parts of the world, and it becomes very difficult.
Part of the problem that lawyers have, too, is that we have very strong ethical guidelines. A lot of that revolves around client confidentiality. Sometimes clients don’t even want the public to know that they’re being sued, even though that’s publicly available. And 99.9% of civil lawsuits settle. They don’t go to trial. So, where is the outcome data? What did that case resolve for? How much did it cost to get to that resolution point? Who has that data? It’s not reported to the public. In fact, it’s subject to a confidentiality agreement and a settlement agreement.
My holy grail, and something I would love to develop, would be a predictive application using analytics that would help use that data. But then, as lawyers, how do we anonymize that data or de-identify it in a way that can preserve our clients’ confidentiality? And then there’s the whole question about whose data is it? If there’s an outcome, is that the client’s data? Is that our data, too, if we de-identify it or we anonymize it? Is there a way to use that data within the constraints not only of our legal obligations but our ethical obligations as well?
Knowledge at Wharton: Anonymity and privacy seem to be on everybody’s tongue these days when it comes to data. Do these concerns compound the challenges?
Walton: Yes, and it’s a shame because the power of legal analytics is immense. People don’t understand how potentially powerful it can be if used in the right way. But we’re all fumbling around our ethical and legal obligations in terms of what we can do because lawyers are risk-averse to begin with. I’m a little bit different. I’m more of an entrepreneurial guy. Our CEO, Michael Heller, is an entrepreneurial guy. But a lot of lawyers aren’t like that. We’re taught to evaluate risk and to guide our clients on risk, so we’ve become risk-averse because we see risk in everything.
When you’re dealing with legal analytics, it’s risky because you don’t want to violate your ethical duties and have an accidental client disclosure, which is a breach of a settlement agreement on a case. But if you just took those concerns out of the equation for a minute, then you could come up with data on case outcomes for different types of civil cases. The power you could have with that data is immense, and we have the technology to do it.
Palantir Technologies can take all the different buckets that the military creates and put them on one platform and then be able to predict with 80% accuracy where the next IED was going to be placed in Afghanistan and Iraq. You can take those similar types of buckets of data from legal, and you’re never going to be able to predict something with 100% certainty. But if you get up into the 80% certainty [level], and you’re predicting the time it’s going to take, and you’re predicting the cost it’s going to take, and you’re predicting what the potential value is of the case, we have all the data streams to do that. It’s just trying to navigate and get that data in one place.
Knowledge at Wharton: From the Wharton Customer Analytics’ perspective, what are the challenges in getting these types of data sets when ownership is not clear?
Iyengar: I think that’s a very good point that Dave brought up. There are lots of questions in terms of anonymity. Let’s start with that. There have been a lot of publicized failures. For example, the Netflix challenge that had come a few years ago. Netflix was releasing anonymized data where you had to predict people’s ratings. Somebody out there took that data, merged it with IMDB ratings data and tried to figure out who these people were. Not surprisingly, Netflix decided to stop that challenge, but that was a debacle that they didn’t want.
“The question for many of these companies is, what are the legalities around it?” –Raghuram Iyengar
There have been many other similar cases. We have to be thinking carefully about not just the type of data you’re collecting, but how people can perhaps put different data streams together, something that companies may not have thought about, something that might be available out there but doesn’t reside within the company. How they can be merged together to de-anonymize has become an increasingly big problem. So, thinking carefully about different streams of data, what people can do with them, how people can in some sense de-anonymize their customers — it’s going to be a big challenge.
The other issue is thinking carefully about what you should and shouldn’t do with the data. In some sense, going back to the ethical guidelines. There have been some cases where companies have been a little too proactive. The Target example is a big one. They have been proactive in using the data to do predictive analytics [for customer purchases]. But where is the fine line between using that data proactively versus “creeping out” the customer?
Knowledge at Wharton: This also touches on human resources and the assessment of job candidates. How does that play into it?
Walton: I think what you’re starting to see our clients do is to take the work that Angela Duckworth did with Grit and develop kind of “grit” algorithms. We used to say, “OK, if a kid went to an Ivy League School, they must be smart and that’s going to increase their chance of being successful. So, I’ll [hire] an Ivy League brand.” Now that’s changed to, “Well, it’s not just your education, it’s your emotional intelligence.” We’ve learned all about emotional intelligence, and how that relates to grit, your ability to persevere through long-term challenges. That’s something that I think law firms are going to [pay attention to] in the future, too.
But there’s a fundamental barrier to law and innovation, which would include [these kinds of] analytics. We’re a professional services organization. A professional services organization has a different economic incentive, a different economic model. We try to cash out at the end of the year. Our firm and all firms carry some money over from year to year for tax purposes, but the way lawyers get compensated is year by year. So, there isn’t a lot of stomach or interest in saying, “I’m going to invest in a five-year program for analytics,” when half of the shareholders in the firm might be thinking, “I might not be here five years from now. I want my compensation now.” Instead of you taking that profit and throwing it into a five-year program to improve the firm as an entity, well, lawyers are jumping ship all the time now. They’re thinking, “No, I want mine now.”
Michael Heller is very adamant that we aren’t going to think like that. But we’re a different firm. A lot of firms aren’t like Cozen. Michael is very entrepreneurial. He thinks ahead. He thinks like a technology guy. So, that’s the first barrier, and we’re still trying to overcome that barrier in the legal field.
“Thinking carefully about different streams of data, what people can do with them, how people can in some sense de-anonymize their customers — it’s going to be a big challenge.” –Raghuram Iyengar
The other barrier we have in the legal field is lawyers think in binary terms — lawyers and non-lawyers. “I’m a lawyer. I know this stuff. I have a law license. I had to pass the bar. I had to go to law school. You’re not a lawyer, so you don’t know.” When lawyers are trying to run businesses and trying to innovate, it’s usually just with lawyers or people who have worked in law firms and have a seat at the table. With our Cyber Solutions & Data Strategies group, I’m trying to build a cross-functional team where we are inviting non-lawyers to the table…. I specifically want non-legal people to have a seat at that table because I’ve been practicing law for 25 years. I know a lot about the law. I need people to help me who don’t know about it. That’s why I try to partner with people like Raghu.
I think I’m the only lawyer who took the business analytics class. I’m trying to go outside of the law to learn about analytics and innovation, which I think are hand in hand, and then trying to bring that back to law based upon what other non-legal companies are doing.
Knowledge at Wharton: How rare is it to have a data scientist on staff at a law firm?
Walton: Very rare. I think your most progressive firms are doing that, but you also run into some internal battles at law firms because lawyers have a tendency to look at IT as being the panacea for everything that’s technology-related, including analytics. Data science and IT — it’s a different skill set. It’s like being a tax lawyer and a personal injury lawyer. We’re both lawyers, but you wouldn’t want your tax lawyer trying your personal injury case, and you wouldn’t want your personal injury lawyer doing your taxes.
A lot of law firms still kind of see IT as the panacea for all this, and they put too much on IT’s plate and get mediocre results. If you’re going to really be serious about using analytics and about being innovative in a law firm, you have to be in the fast lane. You have to be able to iterate over and over again because you have to experiment, learn, pivot. And then that’s how you get ahead.
What we have in our group, I think, is pretty unique. I’ve been told by people at Harvard Law there are only three firms in the country to have something like what we have. But at the same time, we have to prove it. We have to launch products. We have to develop alternative revenue streams because it’s not an academic exercise in the long run. What can we do to develop revenue and improve our client relationships over the long haul?
“If you’re going to really be serious about using analytics and about being innovative in a law firm, you have to be in the fast lane.” –Dave Walton
Knowledge at Wharton: What advice would you give to lawyers who want to start bringing analytics into their practice?
Walton: I think it helped me very much to take the analytics class with Raghu. What I try to do is study books that lawyers wouldn’t read but that people who are in the business world and who are interested in data science would read. I learn from them, and then I try to extrapolate from what they’re doing to see what might work, or not work, in the legal field. I study what businesses do outside of the law because our enterprise clients are five to 10 years ahead of us on all of this.
You also have to partner with non-legal expertise. I’m talking about people who have never worked in law firms. They will see things in your data that you wouldn’t think of because you have a confirmation bias, because you’re already a lawyer.
Raghu and I are talking about starting an innovation lab, or something like that. We could give certain types of data to the experts and see what they can come up with. That’s where you’re really going to have true innovation. You take the lawyers out of it and let the data experts take the first crack at it because they’ll come up with things that you wouldn’t even think about.
There’s not a lot of analytics in the law right now. Data scientists are moving away from Silicon Valley, because it’s so crowded, and saying, “Where are the other opportunities?” I think law is a prime opportunity.