Premonition is an ‘Artificial Intelligence system that mines Big Data to find out which Attorneys win before which Judges.’
Toby Unwin, a Florida-based Brit, is the co-founder and Chief Innovation Officer of Premonition. Toby is a fascinating guy who has undertaken a myriad of roles and adventures in his life (for example, he was inaugurated as the Republic of Austria’s Honorary Consul in Orlando in 2006 and once attempted to finish 3 years of law school in 90 days).
Like Milan (who interviewed Toby on behalf of The Legal Forecast), Toby is a big fan of Tim Ferriss, the prolific US author, entrepreneur and “life-hacker”.
Q&A with Toby Unwin (Premonition)
Toby, what is your background?
I’ve had what’s politely called a “portfolio career”. This means I’m decent at quite a few things, but expert at nothing. I know more about tech than most lawyers and more about law than most technical people. Throw in a lot of entrepreneurial ventures and the kind of mind that always asks “Why not?”. I came up with a question that’s now obvious and much copied – “How often do lawyers win?”. At the time it was heretical.
Please tell us about your time at the University of Southern Queensland!
When I decided to go to law school (around the time I was 40), I decided I didn’t want to take 3 years doing it. I did a lot of research into how to pass a course with the minimum amount of effort. Then enrolled in 5 schools and took the classes concurrently. I passed 3 years of law school in a year with only 90 days of study. I got dropped from a course at Open University in the UK because an admin decided my workload was too high(despite my protestations that I’d already finished).
“You can’t have finished, the course hasn’t started yet.”
“I got a book and read it, like Abraham Lincoln did. I’ll take the exam tomorrow, if you’d like.”
I’d run out of Universities in the UK that let me take individual course, so started looking further afield. University of South Queensland were flexible and good to deal with, so I did e-law there.
After all that effort, I couldn’t combine any credits, so had to do it all over again…. I’ve basically done law school twice.
Specifically, how did you come to be the Co-Founder and Chief Innovation Officer of a legal analytics company that deploys data mining and analysis to determine individual lawyer win rates before judges?
I got sued a lot and it really pissed me off. The United States has no “loser pays” rule, so frivolous lawsuits reign. For example, I got sued for not having an elevator in a 1 story shopping center I owned! The more I learned about law(I’d read over 18 feet of law books before I set foot in law school), the more I realized that many of the lawyers I’d hired weren’t that great and were basically going through the motions. I also noticed that the other side occasionally hired some very aggressive counsel, who were clearly better than my picks despite not having good law or facts on their side. I used to obsessively collect “Top Lawyer” lists, now, of course, I realize they’re total bullshit. There had to be a better way…
I got through law school without using LexisNexis or any of the citation systems. I’d look up a case in Wikipedia and if it wasn’t in there, it wasn’t an important case, so I wouldn’t bother learning it. I imagined these were massive systems that had “all” cases in them. I was shocked to find out there’s practically nothing there and the courts weren’t linked up. After talking to some really unhelpful clerks of court that struggled to understand the concept of the “public” record(I later found out that many courthouse staff have lucrative sidelines selling court data to structured settlement funds, private investigators, etc), I decided the best way to proceed was simply to scrape the courthouse website.
We scraped 1, then 2, then a few, then re-built the system from scratch probably 4-5 times to allow it to scale to thousands of courts rapidly. Court website scraping is standard industry practice that most legal information providers do, it’s just really hard to do at scale and normalize the data. We cracked that nut and left everyone else in our wake. Currently we’re downloading around 8,000 cases an hour from all round the globe. That’s bigger than many analytics firm’s entire databases.
When we got the data, I wrote some algorithms to assign winners and losers, then let the system run. The results blew our minds. Let’s just say justice isn’t blind.
What does a Chief Innovation Officer do?
When Guy and I started Premonition I told him the company would be worth more if he was CEO. He’s an amazing businessman who’s been tremendously successful in a number of different industries. The fact that none of us came from law didn’t deter us, In many ways it’s been an advantage, but I think it probably took us a year more than it should for people to take us seriously. The other day a legal analytics CEO told us his standard counter to people asking about Premonition is that it’s all smoke and mirrors. We offered him a demo and he hasn’t returned our calls since. There’re still people that don’t want to believe, but data speaks for itself.
I like to come up with new ideas and turn them into reality, Guy is the expert on the business side, so I focus on the tech. Chief Innovation Officer reflects this, I think. In the future as we mature, I’d like to do innovation full time.
For those of us who don’t know, what exactly is “big data”, “data mining” and “data analytics”? What are the ramifications for the legal sector?
In law many of the Emperors have no clothes. Big data brings transparency to an industry that’s traditionally had little. Kings will fall, heroes will rise. Nothing will ever be the same again. Neil deGrasse Tyson said “Science is true whether you believe in it or not.” It’s the same for legal data. You may not like it, but it’s here. There are huge opportunities for the first movers because law firms are very difficult to restructure due to all the industry restrictive practices. Those that bury their heads in the sand will wake up finding their clients have deserted them and they can’t afford to feed a hundred litigators “sitting on the bench” with no work to keep them busy. While a boutique firm with phenomenal litigators can now land major accounts because they have the stats to back them up and General Counsel can’t risk low stats of their own by giving “work to their mates”.
Do Premonition’s findings suggest that some advocates perform better in front of some judges? If so, are you able to give us an insight into what this advantage is attributable to?
Absolutely. We’ve found the lawyer/judge pairing is worth 30.7% of the average case outcome. Even our detractors concede that some lawyers are better than others. There are numerous reasons why this might be the case – skill in advocacy, ability to construct a strong argument in law or equity, personality, good looks, who really knows? We try not to theorize as we’re so often wrong if we guess. We do know that these advantages are real. Whether they are fair is debatable. Why they exist?, we’ll leave for others to speculate on.
Some, particularly the public, may find it disturbing that lawyer/ judge pairings alter the outcome of a case – should they be disturbed?
Probably. Justice isn’t blind and law isn’t fair. The only place where the good guy always wins is TV. We’ve found the general public are a lot more comfortable with the idea that some lawyers are “winners” and some aren’t, much more so than lawyers themselves, though most privately concede that relationships matter and “facts and law have lost too many times.”
Eventually, I believe every great change in law will come from transparency. The Heisenberg Observer Theory states that people behave themselves better when they know they’re being watched. As Peter Diamandis says “Most evil is done in the dark.” I’m not saying that the profession is evil, but its incentives are not well aligned with clients. There are few areas that are not improved by transparency and measurement.
Do you think publishing results like this will negatively affect public confidence in the courts and if so what do think the solution to that may be?
In the short term this is quite likely. Many lay people say “I knew it”, or give me a knowing smile when I tell them that we know which lawyers win before which judges. Many general counsel, people who pick lawyers for a living, have told me they’re surprised the personality effect is only 30%. In the long term, I believe that transparency and measurement will increase trust in the courts. Finding your judge decides in line with his peers, doesn’t play favorites and is rarely over-ruled on appeal would be a comforting thing for litigants to know and believe.
Further, does it risk creating a kind of cycle, where clients / solicitors continually select ‘pairings’ with good results and thereby reinforce biases that might exist?
This is a very valid point. I believe it is likely to lead to The Matthew Effect – “He that hath shall be given more”. The flip side is not over paying for poor performing lawyers. After a while market forces begin to take hold though. The top guy isn’t available because he’s too busy, or takes on too much work, so his win rate suffers. This is where other lawyers compete on factors such as case duration and cost. Flat fee litigation, or flat fee per case phase, first filings, discovery, depositions, etc? You’ll see that when the market becomes transparent and competitive. Studies have shown that clients are willing to pay 20% more for flat fee services due to higher certainty on costs. Are lawyers really that incapable of figuring out how long a typical lawsuit will take? I think not.
We read an article about Premonition that described a study which “revealed that the correlation between pricing and performance is weak, and that a law firm’s choice of barrister was 38% worse than random”. Are you able to unpack this for us? How was the figure of “38%” arrived at? 38% worse than random seems nuts!
I agree, 38% worse than random is totally nuts.
That was the win rate: re-hire rate in the United Kingdom. The main reason for a barrister to not be re-hired is failing to get along with the instructing solicitor. Over time this leads to a “hire your mates” system where poor performance is ignored. “Winning isn’t important” is something we’ve heard from more lawyers than we’d like. Guess what? Clients like it – it’s what they pay for. “Charles is a good chap, we were at school together. We always use him for this type of case and he’s got a lovely wife” is the kind of nonsense that clients won’t be tolerating when they discover that the 2nd, 3rd and 5th most hired counsel in the UK Court of Appeals have win rates in the 20s%. The days of “hire your mates” are numbered…
What is the future for Premonition in Australia?
We’ve added the Australian Courts and are interested in finding people to partner with there. We’re firm believers in strong partnerships and having people with the right contacts in the local markets. The insurance companies drive litigation, so we like to find people with strong relationships with insurance claims managers and general counsel.
Will you be hiring and how do people apply?
We’re certainly looking for partners. Most of the people we wind up doing business with were “fans” who reached out and asked how we could work together.
Finally, what is your “legal forecast” i.e. what will the legal sector look like in 10 – 20 years from now?
I think we’ll see business as usual for about 4-6 years. Then there will be a grand shifting of major litigation accounts from firms with established GC/managing partner relationships to smaller firms with proven performers. Brand will suddenly be worthless. Law firms are difficult to restructure and not noted for their agility. Many will disappear. With stats driving litigator selection results, cost and efficiency will suddenly become vital. The first firms to roll out automated contracts, e-discovery types solutions to all areas of their operations, etc will be able to swipe accounts from laggards. This doesn’t happen now, because the legal market is opaque and inefficient. But that day will come.
Source: thelegalforecast.com