Data is the next frontier in just about everything. Every tech company claims to be a data company (or an AI company, which is basically the same thing these days), and for good reason. Data has become extremely powerful thanks to artificial intelligence’s ability to comb through huge amounts of information. That information contains our spending habits, our driving behavior, our responses to brands, and so much more.

It’s no surprise that data aggregation and interpretation has come to the legal market. The truth is that it has been here for years. Insurance companies have been sharing lawsuit data for at least a decade. Now that same data is being mined for more useful information. We don’t only consider how much cases are worth but also information on individual lawyers and judges. There is no doubt that this information will be used by all law firms in the near future.

Data Available to Law Firms Today

There are three areas where data is being used today: case results, attorney profiles, and judge profiles.

The compilation of case results has been happening for decades, but with help from artificial intelligence more cases can be collected and nuanced data can be teased out of the information. No longer are we only concerned with the final amount paid to a plaintiff, now we can identify specifics about the case such as location, medical damages, and the nature of the injury. We can separate a soft tissue case in Kane County from a broken ankle in Cook County.

Some companies claim to be able to rate attorneys based on past performance. Premonition claims to be able to compare attorneys based on win rates, hourly cost, and the average length of time a case is open. Based on those metrics, Premonition ranks lawyers from best to worst. This is primarily offered to insurance companies and in-house counsel, both of which are in the business of hiring defense attorneys.

Premonition also claims to have profiles for judges. This is not unique to Premonition; many companies claim to have data on judges. The theory is that with enough data you can predict what a judge will do on a specific motion. This data is largely limited to federal judges, where the case details are more available. I have yet to see a company offer similar information on state court judges.

How Useful is this Data?

Information on the valuation of cases is very useful, and my guess is that virtually every insurance company uses that form of data on a regular basis. The data is limited somewhat, however, by the availability of information. For example, much more data is available in large urban areas than in rural areas, and the data doesn’t always apply across geographical regions. Chicago juries award more money than suburban juries, and there are perhaps 100 times more cases in Chicago than Aurora (Aurora is the second biggest city in Illinois but has 1/10^th the population of Chicago). The information gained by analyzing Chicago juries will not be applicable to a case in Aurora.

The information supposedly gathered on attorneys is even less useful, in my opinion. They claim to rank lawyers based on win rate, hourly fees, and length of time a case is open. The lawyer has virtually no control over any of those factors. I’m not even sure how a win rate is calculated. Civil cases are never won or lost. Usually, a win is any verdict higher than you were expecting (or lower if you are defense counsel). There are very few cases that return a $0 verdict. Frankly, I think there are plenty of cases where both lawyers win.

Lawyers also have little control over the facts of the case, which are hugely determinative of the end result. Did the defendant rear-end the plaintiff? Pretty hard for the defense attorney to win that case. Maybe he or she can reduce the award, but a complete win is out of the question.

Also, difficult clients frequently dictate whether a case even goes to trial. I will give an example from my career. I handled a divorce case many years ago that had maybe 10 issues for trial. Our client was slated to lose on virtually all of them. The guardian ad litem was completely against her, and the odds of success (however you define it) looked grim. The trial went extremely well. We actually won on 9 of the 10 issues. To the judge and every lawyer in the room, it was a resounding victory. However, my client considered it a loss because we did not win on the one issue (btw, that issue was ranked at “not important” by our client prior to trial). So was that a win or a loss?

Similarly, an attorney has little control over hourly fees. Our market generally defines fees. If you charged drastically less than the competition, clients would be suspect of your abilities. And fees tend to follow experience. A brand-new attorney will have a lower rate than a seasoned litigator (and perhaps a better win rate, too, given the small number of cases). If you rank attorneys by the rate they charge, you will likely hire attorneys from smaller towns with less experience. Those are odd criteria with which to make a decision.

Lastly, lawyers have very little control over the length of time a case is open. There can be delays for all sorts of reasons: difficult clients, service issues, troublesome witnesses, complicated issues, etc. Some lawyers are slow, don’t get me wrong, and they are a headache to deal with, but to rank attorneys by the length of case misses the nuances of handling complex litigation. The lawyer is merely one factor in the length of time a case requires.

Judge profiles are potentially more useful but probably not for specific motions. Does this judge grant a lot of motions for summary judgment? A foundational question must be how many motions for summary judgment does he or she hear in a year. The higher the number of cases, the more useful the data for sure. This data will be more useful for judges in large counties, who handle many cases. A judge in a rural county might only hear 10 motions for summary judgment a year. With a data set that small, outliers are more likely to impact the data, making the information less useful.

The Data Will Become More Useful

We are just at the beginning of this trend in data use. As AI gets more sophisticated and courts store more data online, the information gains will be great. It’s not hard to imagine a world where every decision of a judge is recorded somewhere – in fact, that is already happening. The question is when will our data mining be effective enough to catalog everything. Advances in optical character recognition and natural language processing will usher in a new era of data gathering because machines will be able to understand the handwritten orders found in most courts. More importantly, the data will be organized in a way that is more useful to law firms and insurance companies.

What are the Implications of Better Data Sets?

I had a negotiations professor in law school that had an interesting experiment. He sent half of the class outside and then asked the remaining students two questions:

1. Is the population of Pakistan greater or less than 90 million?
2. What is the population of Pakistan?

Then he asked the second set of students, whom he had sent into the hall, two different questions:

1. Is the population of Pakistan greater or less than 170 million?
2. What is the population of Pakistan?

The first group of students predictably guessed that the population of Pakistan was around 90 million, and the second group put the number around 170 million. The population of Pakistan is 212 million according to the World Bank, but that was not the point of the exercise. The professor’s point was that the first mover in a negotiation frequently dictates the general area of the negotiation.

But this applies to big data sets, as well. If two attorneys have access to that same data which puts the value of their case between $30,000 and $50,000, the case will invariably settle around those numbers. This is especially true if the data set is large and trustworthy. There is nothing wrong with this, but isn’t the data influencing the settlement rather than the settlement influencing the data? Won’t the data become even more standardized? Isn’t this the tail wagging the dog?

Believe it or not, this already happens in workers’ compensation law. Every settlement or decision is recorded and the attorneys spend their time arguing whether the current case is worse or better than comparative injuries from other cases. This will start to become the norm in civil cases, as well.

In fact, the only arguments made by counsel might be to move the case from one data set to another. For example, I might argue that the case should be considered a broken leg rather than a broken ankle because a broken leg is worth more according to the data.

These issues will occur. The only question is when they will become mainstream. The other natural issue will become: what is the role of attorneys in this data-driven process? That’s the question that should make every civil litigation attorney sit bolt upright in the middle of the night.