Wednesday, August 27, 2014

Ravel Law: Legal Research Radically Reimagined

It is an understatement to say that Ravel Law is not your father's research platform.
You are not just using a new product, you are entering the "world of Ravel"... and you need to check all your preconceptions about legal research at the door.

Ravel Law makes me feel old. It has some similarities to Fastcase but also some major differences. Both  products are the brainchild of young lawyers hellbent on reinventing legal research. Both developed their own innovative search engines and visual approaches to  displaying search results.

I have often wondered how young lawyers will understand legal research without benefit of taxonomical hierarchies, digests and headnotes. Only time with tell, but I have to give the Ravel innovators the benefit of the doubt and celebrate something that may "speak the language" of "born digital" generation of lawyers.

The Ravel Research Universe

Daniel Lewis co-founder of Ravel Law grew up in a family of lawyers. When he attended Stanford Law School he became convinced that the legal profession needed a new approach to legal research. He and  co-founder Nik Reed, a fellow Stanford law school alum developed Ravel in collaboration with students from the Design School at Stanford.

Lewis describes the Ravel Law platform as offering a new kind of analysis by using machine computing and data visualization. Ravel built a platform to appeal to younger lawyers. 

For those of us who learned research using taxonomical hierarchies, viewing research results on Ravel is like landing in an alternate universe. Interpreting  Ravel search results requires the learning a “visual language.”  Lewis believes that  Ravel’s landscape -- the visual display  of search results --conveys more information than can be displayed by  the traditional "results list of cases." This may be true but first you have to learn their "display language."
Key to the Ravel Universe:
  • Each circle represents a case
  • The size of the circle represents its importance
  • The Line is the citation
  • The thickness of the line represents the depth of treatment.
At the bottom of the screen you can display the chronology which shows the evolution of the law. The timeline  shows the whole universe of cases related to a search. Filters allow you to limit results by court level. 

The Radical Ravel Leap. Here’s where Ravel breaks with the competition. Ravel only displays the top 75 cases which result from the search. The assumption is  that the algorithm has gotten you the best cases in the top 75. This breakpoint  forces the  researcher to focus on the most highly relevant cases. It also requires complete trust in the search algorithm.The researcher can continually expand and refocus the search to include more cases.

Case with citation history display

Jump Cite Ranking. A "jump cite" refers to a reference to a specific issue within an opinion. The citators on Westlaw and Lexis both have systems for locating cited principles within a case, neither has developed a system for ranking these  specific internal citations. Ravel has developed a unique system for ranking jump cites.

Star Reading system. Ravel highlights the most relevant text in the jump cite in order to show the relevance of a particular page in a particular case. Lewis describes this as  “building out good and bad law on a page by page basis.” The star reading system assigns stars to each page of a case using a 1 to 5 star scale, depending on how many times that page has been cited to by other cases. A one-star page has been cited to at least 5 times, and a 5 star page has been cited to at least 2,000 times. The cases that get listed as citing to a particular page can then be ranked in two ways: 1) by date, 2) by rank (which means how many times that case has itself been cited).  

"Killer Ap" Coming Lewis gave me a preview of an exciting new feature which is not currently offered by any competitor. It is the kind of feature that law firm partners are likely to  want on their desktops. This new feature will  use a "big data"  analysis to  fill one of the voids  where lawyers rely on intuition and anecdote because they are lacking objective data.

Interaction vs Reading. Ravel unravels every preconception you have about legal research. The thing that will most likely  appeal to young lawyers is that Ravel begs you to interact with with it rather than read. The law still requires deep reading, analysing, distinguishing finer points of arguments. Will it seduce lawyers away from reading into focusing on narrower and narrower slices of text? The optimal legal skill set of the future will require both the ability to engage in visual data manipulation and reading full cases.

It is exciting and  also humbling for me to witness the birth of a new generation of legal research platforms. Ravel offers legal research reimagined ... untethered from the inherent constraints as well as the familiar conventions of research systems born of print.... Ravel Law opens up a brave new world of legal research for exploration...


  1. Thank you for this insightful review of Ravel. Ravel works well for simple descriptive searches, such as "fraud on the market" - a search that predictably ranks Basic Inc. v. Levinson as a highly relevant case. (Ravel doesn't yet (?) appear to include among these search results Halliburton Co. v. Erica P. John Fund, Inc., the latest U.S. Supreme Court ruling on Basic, even though it can be found on Ravel.) But questions of procedure involve nuances that automation has yet to distinguish. Consider an example of what Dan Dabney calls "syntactic complexity" :

    “If a person a waives his or her right to trial by jury in one trial, can a jury trial still be demanded in a subsequent new trial of the same matter? The key words for this question, 'trial,' 'jury,' 'waiver,' and 'retrial,' are common in judicial opinions, but discussions of the specific point of law of the question are relatively rare. A computer cannot reliably find cases that are on point because too much of the meaning of the desired cases is tied up in the syntactical relationships between the words, which are not 'understood' by the computer." (The Curse of Thamus: An Analysis of Free-Text Document Retrieval, 78 LAW LIBR. J. 5, 19-20 (1986))

    I tested Ravel to see if I might find relevant cases, including those cited by 5 Cal. Criminal Law § 459 ("Criminal Trial - Right to Jury Trial - Waiver of Right - Effect of Waiver on Retrial"). Upon cursory review, I found none. Of course, one instance hardly supports a generalizable inference. And I agree that such start-ups as Ravel (and Judicata) represent new and exciting models of case law research. But Dabney's example at least suggests that taxonomical hierarchies continue to have unique value.

    1. I think you make an excellent point Michael. I'm curious: do you know of any search engine that does a better job finding your specified cases with these keywords 'trial,' 'jury,' 'waiver,' and 'retrial? And does this query "Effect of Waiver on Retrial" provide a better result?

      I'm trained in civil law so I can not verify nor pretend to be able to validate your findings. Here's my point: if you search for 'apple', how would anybody know if its the fruit or the company if you do not add 'juice' to the query?