Lex Machina: ‘Moneyball’ Meets Patent Lawsuits

By Patrick Hoge. This article originally appeared on San Francisco Business Times, September 3, 2013.

Federal investigators recently called for patent officials to study patent lawsuit trends in an effort to improve the quality of patents, which many critics say have been overly broad and vulnerable to litigation abuse.

To support its recommendation, the Government Accountability Office relied on information from a pathbreaking Menlo Park-based startup called Lex Machina that is using data crunching technology to help lawyers and their clients better predict outcomes and choose strategies.

“We’re defining a new field of legal analytics,” said CEO Josh Becker, a venture capitalist who took over Lex Machina in 2011.

Lex Machina — Latin for “Law Machine” — is a 20-person, for-profit company that was spun out in 2009 from a three-year-old Stanford University Law School and Computer Science Department project called IP Litigation Clearinghouse.

Today, Lex Machina’s web crawlers have compiled data from 147,000 intellectual property cases from 2000 to the present, including 41,000 patent cases, 42,000 copyright cases, 50,000 trademark cases and 11,000 antitrust cases. Where patents are concerned, the technology grabs documents from the courts and from theU.S. Patent and Trademark Office and renders them searchable.

Mining such a trove enables clients to see procedural tendencies of particular courts, histories of attorneys and other parties, and probable litigation costs.

“That kind of data is helping people to change their strategy, rethink how they are approaching different cases and ultimately spend their money more wisely,” said Becker.

Lex Machina announced its $4.8 million Series A round of funding in May and is currently expanding headcount, having just hired vice presidents of marketing and sales. Revenue over the last 12 months was about $2 million, and the number of customers has nearly doubled in six months.

The company came out of a non-profit entity dreamed up by Stanford Law Professor Mark Lemley in 2006 and built with co-founders Joshua Walker and George Gregory. It’s approach regarding patents works for a variety of other types of cases.

“The architecture and processes we have created can be used for any kind of federal court data,” said Lemley, who is on Lex Machina’s board of directors.

Lemley’s intellectual property initiative and others, like the Stanford Center for Legal Informatics, have made Stanford University a center for innovation in applying technology to the law.

Another Stanford spinout founded last year is Palo Alto-based Ravel Law, which uses visualizations, analytics and collaboration tools to aid in legal research. It has raised $1.1 million.

Two Stanford graduates last year were also among the co-founders of San Francisco-based Judicata, which uses “specialized case law parsing technology and algorithmically assisted human review” to turn unstructured court opinions into structured data. It has raised $7.8 million, with early backers including PayPal cofounder Peter Thiel, another Stanford Law grad.

For the GAO, Lex Machina produced data covering 500 patent lawsuits from 2007 to 2011.

Analysis revealed that during the period, the number of defendants in patent infringement lawsuits had increased 129 percent.

Most of the suits were lodged by product makers, but a fifth were brought by so-called “non-practicing entities” — patent-holding entities ranging from universities to licensing firms that don’t make products. Non-practicing entities are frequently derided as “patent trolls,” but they can range from universities to for-profit entities, and Lex Machina calls them “patent monetization entities.”

The GAO also found, however, that lawsuits involving software-related patents accounted for about 89 percent of the increase in defendants, suggesting that the focus on the identity of the litigant — rather than the type of patent — may be misplaced.

“The GAO study is very timely,” Becker said. “We’re excited we can be part of that and provide the data that is leading to this debate.”