This is a live blog post from a private large law firm meeting. The presenter is Zev Eigen, Global Director of Data Analytics, at of Littler. Zev has practiced in law firms and a law department. He then studied for a Ph.D in data analytics. He joined Littler in 2015. [This is a live post, published as a session ends. Please forgive typos or inaccurate reporting.]
His role at the firm is broad and falls into a few buckets:
- Predict events in HR and litigation. For example, in hiring. Zev says that resumes are not at all predictive (except that typos are negatively predictive). The better way to hire is to match characteristics of success on the job to the candidate. “Going to Harvard is predictive of nothing.” Resumes are just a proxy and not a good one. Getting the data, however, takes a lot of work. If you want “good lawyers”, you have to define what good means. Firms must measure the outcomes for which want to maximize. There are over 60 public social media sources that can inform about candidates. In general, avoid proxies (eg, resumes) and go directly to real data. With real data, you can predict turnover.
- Predict outcomes in the legal realm. Littler has CaseSmart for managing higher volume cases. Zev is now building predictive model: length of case, likelihood of dismissal, opposing counsel receptivity to settling.
- Predict legal risks. The firm also helps client predict legal risks. For example, the firm helps clients with California Fair Pay Act. Use data analytics to avoid risk or defend with existing risk
Also damage assessments… the firm internalizes damages computation.
Zev notes that the most successful businesses – Uber, Google – rely on data science. Companies that apply data science on HR front are more successful.
Uptake of data analytics in legal is slow. It does not make sense for law firms to get into predictive science if they compare themselves to other firms. Looking at other firms, there is no impetus to change because you look the same as your competitors. The same was true for taxi companies before Uber. Taxi companies made the mistake of thinking that no one other than other taxi companies would enter the market for private car services. If, however, law firms compare themselves to universe of services that clients can use to improve legal outcomes, then the need for law firms to innovate becomes compelling. Clients will turn to service providers that best solve their real problems.
Entering the market of data analytics is hard because there are not enough data scientists. Even Google has trouble hiring data scientists. Zev emphasizes the big talent gap. A good data scientist is a statistician who is pretty good at programming or a programmer who is pretty good at stats. Zev has strong preference for the former and values other subject matter expertise such as social science or psychology.
What are the tools and infrastructure required? Start small with low hanging fruit. For example, use stats to predict litigation. You can start with 1 or 2 people. Use a program like R, which is free, or Stata, which is not free. Do NOT use Excel. Run if Excel is the only tool.
What can law firms do with their own tool. The problem that data are siloed, even in big companies, and certainly in law firms. Firms should use their data sources to help respond to pitches and make the case why its lawyers are better suited than lawyers at other firms. Also use data to quote flat rates and be sure you don’t end upside down (loss because of fixed fee). So use data to price, pitch… and to staff.
Audience discussion: One in audience shares using stats and a data scientist to understand performance of a big team of document reviewers. Also tried to determine who was best at different aspects of doc review. Same person points out that most law firms are very bad at lateral hiring. There is a big opportunity to improve lateral hiring. Another audience member reports that some firms are using analytics to improve lateral hiring. Someone questions if there is a big enough data set – Zev asks compared to what. [RF: Excellent point – even if data is sparse, what is the alternative to analysis?]
Zev points out that firms have a lot more data than they think. For example, firms with firm-issued mobile phones have detailed GPS data on where their lawyers are. Firms could disclose this and can do this legally and correctly with proper communication. Audience member cites Law360 article about a firm that is making available where its lawyers are at all times. [RF: Mallesons won innovation awards for its PeopleFinder system to connect client incoming calls to most appropriate lawyers or staff as quickly as possible. This was about 10 years ago; see my 2008 blog post on PeopleFinder.]
One way to get started with data analytics: Ignore the status quo. Ask why you are clinging to the status quo. Just because you have done something a long time does not mean its right. Resumes is the best example – ignore them (other than typos). You can approach data science from either problems to solve or looking for opportunities to do better. One example is how much to rely on insource versus outsource of expert services. Audience member suggests one could use analytics to see if a client is about to change law firms.
Zev says law firm clients are increasingly using analytics to select outside counsel. If law firms don’t do the same, they will lose out on new business.
