in Law

Westlaw and Lexis Nexis: Are Your Days Numbered Yet?

During the first day of law school, they lined us up in a conveyor belt like fashion and pushed us through various vendors whose products we’d use and rely on in school and potentially our careers after. Among these vendors were Westlaw and Lexis Nexis. We were given free and unlimited premium access to both of their databases to assist us in legal research and writing. When it comes to legal research, these companies have a near duopoly.

Access to either Westlaw or Lexis Nexis in the real world would cost an arm and a leg for a subscription. Though it’s supposed to be confidential, a Google search seems to reveal that solos and small firms can pay anywhere from $200 to $500+ a month, and I can only guess that the mega firms spend thousands a month. What’s interesting is that the packages they offer are tiered.

If you want something more “obscure” (not that it would be any more difficult to pull from their database), say a law review from a particular season, accessing that content would require you to pay a premium (potentially a la carte from what I recall). This can be such a burden that I’ve worked for a penny pinching solo who admitted to using a free site (filled with ads) to get her cases and citations . We’ve come a long way, and I’m guessing it’s gotten better, but to do the some of the more novel legal research like trying to admit MRI scans into court as evidence for the veracity of a witness’ testimony (I ghostwrote the underlying motion that got the court to agree to hear the evidence as a newly minted and underemployed attorney in 2009 for $15/hr. Imagine my surprise to see the expert who dropped off this New Yorker article and boxes of CDs filled with scientific articles for me to read through on Mythbusters getting busted on TV several years later), you’ll need access to the good stuff. If anyone is wondering, I believe the defendant in the MRI case backed out and decided against proceeding with the evidence – though I was really out of the loop, so I’m not sure what happened.

To leave us shaking in our boots, our legal research and writing professor told us that if we don’t do well in her class and in school, we’d end up in Minnesota in a windowless room reading new case law and writing summaries for Westlaw as a research attorney (it isn’t true after all, I went in for an interview late 2008 in San Jose at a Westlaw subsidiary, and they had windows). Westlaw and Lexis Nexis are useful to lawyers in two, maybe three ways. 1) You can look up a case’s issue and/or facts by natural language (also through a SQL like query though I don’t believe I was ever that technically minded then), 2) They provide a decent synopsis at the beginning of each case (our professors would warn us against relying on any of that, because it can easily be wrong or partially incorrect), and 3) Something called shepardizing – its namesake comes from some other publication, but basically it’s the process of determining whether a law is still good, if it’s been overturned, or conflicted by subsequent cases that may or may not have authority over that particular court’s holding. For programmers, it’s like getting a compiler warning if something is deprecated.

So was Westlaw or Lexis Nexis good? It wasn’t bad. The natural language certainly didn’t work as well as Google, but almost always, I’d end up finding what I needed (if it truly existed) if I looked hard enough.

When Google Scholar got in the game, things got a bit interesting, but it never got true traction because of its inability to self shepherdize. Nonetheless, the natural language on it was great, and the cited by feature has come a long way since it first started.

So when I first heard about Casetext, I was excited. It was started by young law grads from the two top schools in the country who decided against slaving away at a firm as an associate (If my memory is off, my apologies). Though I wasn’t really actively practicing anymore, I thought that the ability to annotate case law like you would on Genius (formerly known as Rap Genius) is a brilliant idea! From a business perspective, it’s great because instead of paying an army of research attorneys, you can count on contributors to help annotate cases. From a user perspective, it can be like Stackoverflow, where you can potentially find your answer if you’re absolutely stumped or need a bit of clarification. Moreover, the value of what you’re getting as a consumer would be greater than what Westlaw or Lexis Nexis could provide – the analysis. That’s right, the motion or memo would write itself!

When I first got on Casetext earlier tonight, I saw some hiccups on the front-end while trying to filter cases out. Being the narcissist I am, I looked up one of my own appellate cases People v. Powell, 214 Cal.App.4th 106 (2013) by first searching via natural language the case name. It returned thousands including ones from other jurisdictions, so I filtered California and that’s when the hiccup happened. The site froze, and when I returned to attempt to replicate the issue, it worked just fine. Looking at their job postings,  I see that they’re using Angular JS, but I couldn’t really form an opinion because I’m still a front end dunce even though I was recently thrown into a ReactJS project super-part time at my day job. I retried the search, and this time, I used the citation, and it returned nothing matching the case. Bummer. So I’m wondering now if they just don’t have everything, if I lost my touch, or if their search is off?


I tried yet again, but this time, I spaced out the citation between the district, volume, Cal and page number because maybe it matters? Nothing. I actually end up getting back cases from the US Supreme Court. Clicking on the “Criminal Law” filter with one case flagged on it gave me something from Texas – I struck out. I suppose I can officially call it that Casetext just doesn’t have my case, but I didn’t expect it to return things out of jurisdiction when I provided an actual citation. Whether Casetext purports to have all state level cases, I’m not sure, as the site is silent on that.

So I’m now seeing that NLP (Natural Language Processing) as confirmed by Quora, is hard, meaning that Westlaw and Lexis Nexis is actually pretty decent?!

So how are the annotations? I poked around some, and of the ones I saw, they were a bit underwhelming:




The cases are mostly barren, and the ones with annotations aren’t particularly insightful like the earlier ones from when Casetext was first announced. This problem, I absolutely understand, given that it’s a relatively new service and the amount of case law out in the wild. Having meaningful annotations for a substantial amount of cases will be a challenge compared to a discography of 90s era rap music where the number most influential tracks can potentially be narrowed down to 50.

So where do we stand? Westlaw and Lexus Nexis definitely has real competition as more entrants gear up. My money is still on Casetext to do some damage given that easy access to annotations from practitioners and academics are important to legal research. Another startup named Judicata, which is still in stealth mode, could be very impactful as well. Co-founded by Blake Masters (whose name is familiar because he co-wrote Zero to One, and took the notes in Peter Thiel’s start up class which the book was based off of. Sidenote: Mr. Masters, like Peter Thiel and Alex Karp of Palantir, are all Stanford Law grads), Judicata’s splash page suggests they’ll be doing some very fancy NLP work by “Mapping the Legal Genome”. Looking forward to checking it out.

From how things are price-wise in software these days, I can surmise that lawyers won’t be paying an arm or a leg for these new research tools (Casetext is actually available to the public for free). Moreover, these guys can probably make the revenue in ways that Westlaw and Lexus Nexis can only dream of when they’re on an acid trip, or perhaps not make revenue at all, like Russ Hanneman suggests on Silicon Valley yet still headed toward the three comma club. It’ll be exciting to see new start ups come into this part of the nook.

So to answer the initial question, not yet, but soon.

Interested in leaving law to code? Check out my book Quit Law and Code to get started.

Write a Comment


  1. Casetext CEO here (also posted this comment to HN). First, thank you for writing this, William. Hearing honest feedback from people who try out our service and improving is precisely how we’re going to get better. And I love that you share both our enthusiasm for the idea and our optimism for the future of legal research.

    I wanted to note that we will be getting a lot better. Keep checking in over the next few months, and expect to see a lot of major improvements on the issues you noted as well as many other things as well. Until very recently we were a very small operation, and as it turns out, building a competing product to LexisNexis and Westlaw is exceptionally hard. (I don’t want to make excuses – trust me, nobody wants Casetext to be much better more than I do – but the reality is that a perfect product couldn’t be built overnight.) That said, with the backing of amazing investors [1], we just assembled a killer team [2], and now have the people and resources to get there.

    Another commenter (dmix) put it well: “Building hard things takes a long time.” Think of how long it took for a lot of the best sites on the web to become what they are today. We’ll get there – I promise.

    Finally, we do have the case you were searching for [3]. Why it didn’t show up in search is something we’ll be looking into this weekend.


  2. I’m surprised that Fastcase and other Lexis/Nexis/Westlaw competitors aren’t mentioned. Is Shepherdizing still relevant when you can do a full-text search for the citation on Google Scholar? It seems to me that the issue is whether or not Google stays up to date or not.

  3. More importantly, the content that Westlaw and Lexis/Nexis resells is public domain information paid by the US taxpayer. They fought viciously to prevent the government’s own FLITE database of caselaw from being released as well as the DoJ’s Juris database, which contains the same public domain material. It is a scandal that these companies can get away with charging so much for laws paid for the public and whose content the public is charged with knowing.