Lex Machina uses predictive analytics to harness data in IP cases, helping lawyers create litigation strategies
By Sean Doherty. This article originally appeared on Law Technology News, April 11, 2014.
A task legion to intellectual property lawyers is how to develop a litigation strategy that factors in all the complexity of a typical IP case, including current trends in specific jurisdictions and general leadings of jurists. Enter data analysis tools such as Lex Machina (lexmachina.com) that help lawyers craft a successful plan and perhaps even forecast the case result—no longer via a wing and a prayer, but by the numbers.
Lex Machina, born at Stanford University and now backed by venture firms and angel investors, is a web-based software and data analysis service that creates structured or tagged data sets of judges, lawyers, parties and patents gleaned from case dockets and documents obtained from the Public Access to Court Electronic Records system, the U.S. International Trade Commission’s Electronic Documents Information System, and the U.S. Patent and Trademark Office. LM will soon add proceedings before the U.S. Patent and Trademark Office.
All sources are processed by a natural-language processing engine called Lexpressions, which cleans up data (e.g., makes consistent case and party names), codes or classifies it, and then tags it for searching and analysis via LM’s web application.
The web application helps users quickly view the litigation history on a patent portfolio to find asserted patent claims, findings of infringement, claim invalidity, patent unenforceability, and damage awards. Case history and past precedent data lends itself well to litigation strategies, especially if you can predict the parties, the court, and even the judge assigned to the case.
LM’s home page provides a view of all the IP cases in the system, filed by year (since 2006). A search box helps users find cases, case dockets and documents, as well as documents in investigations before the ITC.
The home page has navigation links to the subpages addressing courts and judges, counsel, parties, cases, documents, patents and ITC material.
I clicked on the patents subpage and entered Java-related patents owned by Oracle Corp. (Patent Nos. 6,061,520; RE38,104) and recently asserted against Google Inc. LM does not find information on patents unless they have been asserted in a court action or subject to an ITC investigation. For example, LM had no information on Recommind Inc.’s Patent No. 7,933,859, covering systems and methods for iterative computer-assisted document analysis and review.
LM found one action for Oracle’s Java-related patents: Oracle America Inc. v. Google Inc., N.D. Cal. 3:10-cv-03561, filed Aug. 12, 2010, and terminated on June 20, 2012. The system provided an abstract or summary of the asserted patent, similar patents, and the patent finding of noninfringement and the outcome of the case: consent judgment. Case information included data on the parties, judges, court, all the asserted patents, and the docket entries and documents to the case. The docket entries could be filtered by tagged information, which included a claim construction order, pleading, complaint, answer, jury verdict and judgment.
The Oracle America case data included tagged information that linked to more content on counsel, parties, courts, judges, etc. For example, an outside party to the litigation included Motorola Mobility Inc. One click on Motorola yielded a list of links to all cases where Motorola was a plaintiff (14) or defendant (86). A click on the presiding judge, William Alsup, showed that he has been assigned only 17 of the 323 IP cases recorded in LM for the U.S. District Court of Northern California.
Patent data included a link to a Patent Portfolio Report that lays out links to patent numbers and titles, and the cases where they had been asserted, with findings for infringement, invalidity, unenforceability and damages. The portfolio information is useful as a forward-looking mechanism, too.
From the patents page, I searched for IP related to a virtual machine in hardware that supports Java bytecode translation into native CPU instructions. LM showed five patents asserted in 14 cases to research, which could help me fashion an IP strategy to avoid litigation if I was in development, or engage litigation if I had a patent to assert.
On the home page, a single click on a number representing the total IP cases filed in a year generated a list of cases, sortable by most recent docket activity, earliest filed case, most recently filed case, or earliest/latest terminated case.
The case list can be filtered by numerous facets associated with tagged document content. From this 10,000-foot view, you can select open or terminated cases, and can filter IP cases by type to include/exclude patent, trademark and copyright cases. There are filters for cases with declaratory judgments and appeals, and whether the case went to trial, viz., a jury or bench trial.
Other filters include date ranges for filing, termination, trials, last docket, courts and judges.
Lex Machina’s search interface can be as easy as Google, but also supports custom Boolean syntax to construct targeted queries. It supports Boolean operators AND, OR, NOT with grouping symbols: quotation marks for phrases and parentheses for complex Boolean strategies. For example, a search for the word Java AND (“hardware accelerator” OR “bytecode translation”) retrieved 186 utility patents and two reissued patents. Wildcard searches for one (?) or more (*) characters are supported as well as fuzzy, proximity and regular expression searching.
COURTS AND JUDGES
The landing page for “Courts and Judges” displays top court filings. It was no surprise that the U.S. District Court for the Northern District of California and Eastern District of Texas had the most filings (as of March 11). Thumbnail bar graphs for each court show how patent, trademark, copyright and antitrust cases cumulated from 2010 to date.
You can drill down into case types (e.g., patent or copyright). You can view a list of all cases in the system, filtered by court and case type, which can be further narrowed using filters for case status, date ranges and judges.
The judges section also features the top judges for IP case filings. Leading the pack on March 11 was Judge James Rodney Gilstrap (Eastern District of Texas); Colorado and Delaware vied with Texas for top spots. Search boxes dedicated to courts and judges rested at the top of the landing page.
Handy links help users search for recent documents regarding case transfers to other jurisdictions, judgments, injunctions, dismissals and orders. I requested a list of orders issued in the last 30 days, with the most recent order on top. Clicking on any order sent me to the particular case docket and filing.
Using the search box for judges, I started to type “Koh” and selected “Lucy Haeran Koh” from a type-ahead list that matched the typed characters. Koh’s landing page offers information on the number of IP cases she has handled vis-à-vis district’s other judges. Using tagged data links, I viewed analysis of her cases from filing to termination.
LM takes every opportunity to deliver useful data. It provides judges’ biographical and tenure data; and information on IP cases they presided over (case lengths and outcomes by type: claimant wins, defendant wins, settlement, or procedural termination).
Users can set alerts for new cases and docket entries in courts and before judges. Patent portfolio reports can be downloaded in Excel format. At this time, there is no way to save searches or reports, or return to earlier research.
LM’s annual subscription pricing varies depending on the organization, ranging from $35,000 to $100,000. Introduction pricing starts at $1,000 per month. There are reduced rates for solo practitioners and small IP departments. Federal judges, universities and Congress use it for free (with some caveats).
With Lex Machina, legal professionals can create an IP strategy for a client, or a roadmap for litigation, in much less time than it would take a research professional to amass the IP data for review. It provides easy access to a wealth of IP data, which will yield a fast return on your investment.