The Tyranny of Hunches: Using Analytics to Give Your Firm a Strategic Advantage
Law firms are amazing organizations. They are full of very smart people; they utilize incredible amounts of information, and, as businesses, they generate a lot of money. And yet, for many client decisions — including some bet-the-company decisions on behalf of clients — the most important source of information for law firms is the hunch.
After practicing law for a short time, I became a business owner, and now our company employs four law firms to do our legal work. As a consumer, this hunch-based mode of decisionmaking drives me crazy. I would wager that many clients feel the same way about hunches.
Experience — as a Hunch
How much is my case worth? Where is it most advantageous to file suit? What is this judge like, and how is she likely to rule in this case? How long will it take to reach a disposition? What is my likely exposure? What will the legal fees be? Should I accept this settlement offer? How many important precedents exist, and what’s the most important one in this case? Is this contract clause standard?
These are not abstract questions. They are core to the practice of law, and for clients, these are the major questions they face about legal services, and in some cases, about the very future of their business, their family, or their medical care after an accident.
We answer these questions with hunches. Usually bad ones.
“How much is my case worth?” is not a question that is answered with a number. The answer is a distribution of outcomes from similar cases. “What is this judge like?” is not a collection of war stories, or a question about who threw a fundraiser for the last election. Many friends of mine became lawyers so they could be safe from math and science in school, and they have stayed that way for their entire careers. The bad news for them: This mode of delivering legal services is doomed. The day of hunches is ending.
The Rise of Legal Analytics
Legal decisions in the future will be made with data. Not all at once, but starting now, and increasing every year from here forward. This is not controversial — it is malpractice to think otherwise. Data transforms industries. transforms, I mean destroys. A good example of this is Yahoo!, which was the original way we found information on the early World Wide Web. When a lawyer wanted to find a piece of information on the early Web, one would look at the outline index on Yahoo! — which hand-categorized each page on the early Web. This mode of finding information died completely in 1998, when Google introduced its search engine and the first PageRank algorithm, which put the best result at the top of the list. What was manual, anecdotal, and editorial became automated, scalable, and much better — using not hunches, but citation data.
Data and analytics similarly have transformed many industries, from stock trading to medical diagnostics, from Major League Baseball to Jeopardy. There is absolutely no reason in the world to think that law will be different. Data will eat your hunch for lunch.
Here are a few examples of companies that are pioneering the use of data analysis in law:
• KM Standards — If a client brings an employment agreement to your firm, you might have a very experienced partner to handle the matter because perhaps they have negotiated a few dozen employment agreements. Kingsley Martin and his team at KM Standards downloaded 250,000 employment agreements from the Edgar database and machine-analyzed them to identify which clauses are standard and which are negotiated. A few dozen contracts are experience, but 250,000 contracts are data. KM Standards can tell what the range of salaries is or instantly identify which contract clauses are nonstandard using data.
• Lex Machina — Do you know the judge in your case? For patents, Lex Machina does. The company mines data from IP litigation and analyzes outcomes for each judge. Knowing your judge does not mean that you gave her money in her election or play golf with her — it means that you understand the range of outcomes from similar cases in her court. The company’s start is in patents, but we can expect to see litigation data mining in more verticals soon.
• LexPredict — Courts are random, right? Not according to LexPredict that is the brain child of law professors Dan Katz, Michael Bommarito, and Josh Blackman who are using computer models to predict court behavior. LexPredict analyzes past court cases to derive patterns — but it also crowdsources expertise, hosting an annual Fantasy SCOTUS competition (think fantasy baseball for Supreme Court geeks). LexPredict uses both computer modeling and crowdsourced expertise from Fantasy SCOTUS to predict how courts will decide cases in the future.
This is not an attempt to showcase or endorse these companies. It is simply a look at what is coming in legal analytics. These examples illustrate a larger story: predictive coding for e-discovery, algorithmic stock trading, self-driving cars. Many of the tasks we have done in the past with instinct and brute force, we are now doing better with data.
Analytics in Legal Research
We are doing some of this work at Fastcase. Legal research is the ultimate hunch-driven legal analysis. What is the law of the case before this court? What is the seminal case, and what is the case most like mine? Have I finished my research? Usually these questions are answered with educated guesswork in the form of keyword searches — hunches.
Blue Hill Research has calculated the cost of doing research with legacy tools in a 2014 study. According to Blue Hill Research, the average associate spends 193.3 hours of research time that is written off, at an average cost of $61,856 per year. This is a huge, expensive problem.
The Fastcase legal research system has been a free benefit to members of The Florida Bar since 2005. We use citation analysis, data visualization, mobile apps, and big data analysis to help lawyers identify the seminal cases in their area — to find them fast and with confidence. Forecite uses patent-pending citation analysis to find cases that keyword searches miss. Fastcase 7 includes new analytical tools, such as an enhanced Bad Law Bot, the world’s first big data tool to identify negative case history, and a tag cloud that identifies legal concepts hidden in search results.
Some people call this the Cambrian Age of legal tech, with a host of new tech companies crawling out of the primordial soup onto land. But it is more like the Mesozoic Era — we already have big legal tech companies. They are huge and ferocious with sharp teeth. There is also a new era of legal tech companies on the rise; the companies are nimble, and have some adaptive advantages in data analytics.
So the bad news is that most lawyers are pretty bad at math. But the good news is that most lawyers are pretty bad at math. That means that you don’t have to do very much to be better than the average. Data is no substitute for great lawyering, but great lawyers are masters at using the best tools of the practice. Those tools will increasingly be using data and analytics in the future. Great lawyers with data will always beat lawyers without it. It’s the Mesozoic Era of legal tech. You’re either with the dinosaurs or the mammals, but you have to choose.
Ed Walters teaches the law of robots at Georgetown University Law Center, and he is the CEO of Fastcase, a free legal research member benefit for members of The Florida Bar.