Big Data is all the rage and in a couple of weeks I will be speaking to a large group of risk managers about why we need to be using more of it in risk management. Which got me thinking. Why aren’t we using this more when we evaluate the performance of our attorneys in claims management?
In the book “Think Like A Freak”, the authors speak of their previous research on sumo wrestling and how they determined if matches were thrown. They paired the matches, scored the records and then ran analysis to determine if sumo wrestlers actually threw matches (cheated). Their findings may surprise you, so I will leave you to read the book if you haven’t. What was more interesting was the types of things that the Freaks looked at when trying to determine matches and how we might apply those factors to our attorneys’ outcomes.
Our old stats aren’t enough
Traditionally when we look at our attorneys performance in claims management terms, we tend to run the cursory central tendencies measures (mean, mode and median) and all the fun frequencies – number of wins, number of losses, number of cases, cases by dollars, etc. And while that is mildly useful, if we really want to determine attorney performance, then we need the power of Big Data.
Big Data is all we need
Big Data goes beyond the simple metrics we normally use to look at our loss runs and it tries to provide correlations that we might otherwise never think of. In our sumo example above, the authors looked at the match-up, something that we typically don’t do in claims management. Meaning what was the outcome of Attorney A when the faced Attorney B? What about when Attorney A faced Attorney C. What happens when Attorney B faces Attorney C? .
And while we can probably look at these match-ups using simple spreadsheets, we need the power of Big Data to dive deeper. For example, what happens when we look at Attorney A, Attorney B and Judge A. And what if we throw in the type of claim/case it is? And what if we throw in the gender and race of the plaintiffs? What if we look at the original demand? The terminology utilized in the initial claim? Number of attorneys in the firm? Case load of the attorney?
The list goes on and on. And that is the point that only Big Data can sharpen. Big Data looks for correlations not causation. It helps us predict the outcome of that which we analyze. In this instance, it can help us predict whether Attorney A or B or C will win a case. It is in this knowledge that allows to not only judge the performance of our attorneys but also help improve the performance of attorneys and our litigation strategies. If we know C always beats A and A always beats B and B always beats C, then when we are faced with C, we should certainly assign B. But, leaving only ourselves to blame, we often select an attorney to represent our cases using our failing memories and personal preferences. We often don’t harness the power of Big Data (using the numbers) to help us win.
Next time look at this
The next time you want to judge the performance of your litigation strategies, keep these things in mind to analyze in addition to your traditional stats:
- The match-ups
- Previous match-ups
- Case loads
- Firm type
- Original demand
- Jury or not
- Plaintiff demos
- Case type
- Dollars spent
- Distance to court
- Cases before the court