Data analysis exerts a powerful, often unseen, effect on our lives. Organisations ranging from sports teams to supermarkets employ data analysis as a means of gaining a competitive edge. Moreover there is plenty of data to analyse. Whether it be player movement on a pitch or the swiping of in-store loyalty cards, it seems the rising tide of data is unstoppable.
Law firms cannot hold back this rising tide. Like everyone else, they need to take data, and its analysis, seriously. More particularly, they can start to deploy data analysis to help their clients and deliver their services more effectively.
I have a background in computing, with particular reference to Artificial Intelligence (AI) and legal expert systems. Some of the very latest developments in the field of legal data analytics use some nifty AI technology (see below). I must confess though, my recent interest in data analytics has been piqued by sport. I love sport, and professional sports teams are increasingly turning to data analytics to help improve performance.
Data Analysis in Sport
Appreciating the finer points about baseball is still beyond me, despite reading, and hugely enjoying, ‘Moneyball: The Art of Winning an Unfair Game’ by Michael Lewis (also made into a film starring Brad Pitt). The fascination with Moneyball lies in its explanation of how the general manager of the Oakland A’s, Billy Beane, applied data analysis. He built his teams and traded his players relying on data. More accurately, he focused on particular types of data which had until that time been ignored by baseball coaches. As a result, the Oakland A’s were remarkably successful. The role data analysis played in this success was noted (somewhat belatedly at first) throughout baseball and in other sports. Data analytics in sport is now big business in its own right, being applied to an ever increasing range of sports.
What might be called the traditional coaches in baseball were highly sceptical of Billy Beane’s approach. I am sure that many (if not most) lawyers are equally sceptical about the relevance of data analytics to their work. Not least because, as Ron Friedmann commented in reply to one of my earlier posts, lawyers do not like working with numbers.
In sport, the sceptics were won over because they realised that analysing data properly can provide a competitive edge. I’m sure the same will happen with lawyers. As with professional sport, those applying data analytics properly are likely to emerge as clear winners, forcing others adopt similar methods in an effort to catch-up.
Key Performance Indicators for Law Firms
About 10 years ago, when I worked for Axxia (which was bought by LexisNexis), the company created a product called AxxiaKPI. This presented financial key performance indicator (KPI) data to lawyers, partners and management staff in dashboard format. Surprisingly, to me at least, most law firms at that time appeared reluctant to engage with their data fully.
I’m sure the existence of dirty, and therefore unreliable, law firm data was (and still is) used by some as an excuse not to make use of their data. But dirty data can be cleaned, rationalised or otherwise compensated for.
Law firms hold lots data which can be mined for really useful business development information. Data cleansing and rationalising may appear costly and time consuming, but the benefits of cleaning and mining data should be obvious, especially in the current climate of downward pressure on fees and the oft cited demand of clients to provide ‘more for less’.
An interesting example of how some law firms are using data and thinking statistically appears on the website of Bott & Co, a leading firm of personal injury lawyers. They have made available their case ‘profitability indicator‘ to help other firms clarify their decision making. To make use of the profitability indicator firms need to input things such as cost of case acquisition and processing and average revenue per case. This data – and much more – will be in law firms’ practice management systems (PMS) and can be extracted easily, usually by using built-in PMS functionality or perhaps by using add-on reporting software. The point is, the data is all there and law firms should be using it much more than they were 10 years ago.
Bench-Marking Law Firm Performance using Big Data
It is now possible, at least in the USA, to subscribe to services offering comparative bench-marking based on data from a sample much wider and richer than that which can be gleaned from a single law firm. For example, the most well-known bench-marking service provider, Tymetrix, say they perform their comparative exercises based on billing data from over 17,000 law firms.
The era of ‘Big Data’ has arrived. Data analytic software is now able to apply clever technology, often incorporating algorithms and techniques associated with AI such as natural language processing, to very large data sets. This has the potential to affect, if not revolutionise, the practice of law in many ways.
Discovery and Predictive Coding
Most litigators will know of predictive coding and its application to e-discovery. Predictive coding is where software ‘learns’ from lawyers. Lawyers mark up samples of the electronic documents at issue. These marked up documents are then run through software, which ‘learns’ from the mark-ups. Once the software has done the learning, masses of documentation subject to discovery can then be processed by computers at much greater speed (and consistency) than lawyers and paralegals. The cost implications of applying this kind of technology are obvious. There is much less need for teams of relatively expensive lawyers and paralegals reviewing lots of documentation when the same task can be done better, and more cheaply, by software.
Predicting Case Outcomes
It does not take a great leap of imagination to appreciate that if computers can learn from lawyers sufficiently to categorise documents successfully, they could perhaps do the same to predict the outcome of the litigation itself.
Potentially, litigation lawyers will in future be able key in material facts of a case and then let software run those facts against a huge set of legal case data to arrive at a statistical (probabilistic) estimate of success. This is more than a flight of fancy – see for example the work that Professor Daniel Katz and his colleagues have been doing in this area.
Legal Service KPIs
Although fascinating, the possibility of computer aided case prediction lacks immediate relevance to practising lawyers. For legal work (as opposed to managing legal businesses), practising lawyers can perhaps make more tangible progress by considering what kind legal KPI’s are of interest to their clients, and then collecting data based on these KPI’s.
Axiom has done some interesting work in this area (see Legal Briefing Magazine, June 2014 pp 12). For example, it collects and analyses data about which contract clauses are most frequently subject to negotiation.
It is easy imagine what else might be done in a similar vein. Wouldn’t it also be useful to know which clauses where subject to litigation most frequently? Once this is known, it should be possible to estimate the risk of litigation associated with particular clauses.
Once you start thinking of questions such as these and identify data that can provide answers, all kinds of possibilities present themselves. Paradoxically, this too can be a problem. Only the most relevant possibilities should be pursued. As W. Edwards Demming noted:
If you do not know how to ask the right question, you discover nothing
In the era of big data and legal KPI’s perhaps this is where lawyers’ expertise can be used best, by asking legal data sets the right questions.