Building LawStar
Hi, I’m James (in some circles better known as Mackey). In my career, I am a software engineer that likes working in startups. I regularly write about the things that I am working on, but this is my first time posting publicly.
This article is about my journey starting my own company, selling it a year later, and what I learned along the way. I hope that it’s useful for anyone thinking about how to build products or start companies.
Finding and Validating the Bluebooking Problem
Career choices while you’re in your 20s should be optimizing for learning. Before LawStar, I worked in engineering management at a high growth startup. It was a thrill, but in my last year there, I wasn’t learning much new. So I began exploring a dream of mine - starting a company. I believed I had enough experience to be successful and could re-accelerate my learning curve.
My first instinct was starting a classic VC-backed tech startup, and with the trajectory that LLMs were on when I started researching, I thought legal tech was a good place to start given the volume of text involved in the law. I started pitching some ideas to my friends in law. What I heard a few times was, “that idea sucks - what you should really look at is bluebooking.” Honestly, I’d never heard of bluebooking before, but decided to find out more.
The Bluebook outlines a citation standard similar to MLA or APA citations you might be familiar with from writing essays in high school, only it is much more prescriptive: it is 500 pages long, with minute changes in formatting for every possible form of citation. Legal academic journals, professors, and most courts require citations in Bluebook format. Some law students spend up to 10 hours / week doing bluebook citations, so I started speaking to them.
Second year law students in law journals (like scientific journals, but for legal papers and run by students) often spend hours per week editing the citations of the papers sent to these journals for publication. I surveyed 25 of these students, and a majority of them said they’d pay > $10/month for a service like LawStar. I figured if I was able to successfully target these students specifically, I’d have my little niche and could grow from there.
A quick market-sizing exercise: there are ~110k law students in the US, ~25% of whom are in journals, and half of those are 2nd years. If I could penetrate even 10% of that market at $10/month, I’d be at $165k in Annual Recurring Revenue (ARR).
Weighing My Options
Given the numbers above, I didn’t see LawStar as a VC-backable business. I thought that there was a very good chance that I would be able to get to $100k+ ARR, and a chance that I could get to $1m ARR, but not to $10m. And I was worried about staying excited about LawStar for 3+ years. As exciting as citations are, I thought that I would be more productive and happy working on something that I felt was more impactful and interesting.
That said, I was ready to leave my former job. I had some savings, a clear market, and thought this was a low stakes opportunity to learn and show off my ability to execute. There were 3 outcomes I imagined, in order of preference:
The business works! Find some acquirer for either cashflow or the technology that would allow me to take the credibility and experience from this and take on a bigger problem.
The business works! And opens up a pathway to building a way bigger business in the legal tech space.
The business doesn’t work - go get a job.
I decided I was excited enough about that framing to give myself one year to try to make it work.
Legal Citations Market Landscape
There are a few websites that do what LawStar planned to do. Generally they fall under 3 categories:
General citation websites
The players are sites like citation machine, citethisforme, citethis.net
They offer a bunch of different citation formats, and include the bluebook in that, but their bluebook format is very imprecise. Nobody uses these.
Bluebook citation websites - websites that were built specifically for bluebook citations.
citeuslegalus.com and bestlaw.io were decent websites that got a decent amount of use, but have fallen into disrepair as the builder has focused on other things.
legaleasecitations.com does largely what LawStar intended to do but without automation.
Legal research platform - where users are already doing their legal research. Mostly looking up journal articles, laws, and court cases. There are a few online databases people do their research in:
Westlaw and LexisNexis - these 2 are the 800 pound gorillas in the legal tech space. They made close to $5 billion between the two of them in 2022. They have one-click citations on their sites, but they are often inaccurate.
Bloomberg Law, Fastcase, CaseText, CourtListener, HeinOnline - Seen them pop up, and a few but not many students use them. Fastcase, CaseText, and CourtListener are interesting because they were founded in 1999, 2013, and 2009, respectively, so their founders are still active in legal tech. Fastcase and CaseText were recently acquired.
Building
In two months, I learned how to interpret the Bluebook, plan out what features I’d need, and build the initial framework and first feature - court case citations.
Court case citations were initially all manual input - so a user would need to add all of the information that would be necessary to perform the citation. I showed this to a few law students, to luke-warm reactions. It wasn’t making the citation process faster.
So I built a Chrome Extension that would add a button to one of the main research sites (Westlaw). Clicking the button redirects to LawStar with all of the required information from the web page to complete the citation, and the user would have the full citation ready. This got students excited. It also gave me one of the key insights for LawStar - most users care more about making it easier to create citations than the accuracy of the citations. Both are necessary, but I should focus on speed.
I built out 3 more citation types - internet pages/articles, periodicals, and books - and was ready to launch. For pricing, I benchmarked against legalease citations, which I mentioned above. They were charging $20/month. Because my service was very incomplete, I set the price at $6.25 / month or $60 / year, with the idea that I could upsell or increase the price later.
Launching
I emailed all the people that took my initial survey on October 8th and gave them full or big discounts ... but almost nobody signed up. I knew I needed users to get feedback, so I made it completely free. Even then, people weren’t signing up much. I printed out posters and fliers and posted them up all over Quinnipiac and Yale Law Schools. This got 4-5 more signups, but there was hardly any usage. Clearly something had to change.
I made 2 big changes:
I could see that a lot of people were interested in seeing what the product looked like.
Many people were pressing “Try the Product”, but the first thing that popped up in the web app was a signin modal, and it seemed like users didn’t want to give away their email if they didn’t know what they were getting. I thought about removing the Sign Up altogether, but felt the information that I could get from it (email, first and last name which ) was important for my understanding of who was using and finding success with the product. Instead, I added screen captures of LawStar in action to the landing page, and landing page conversion went up.I built a new feature to automate internet citations. Users could simply post their url into a text box, click “Cite,” and LawStar would parse all the information necessary to complete the citation.
The weekend after finishing these two things (November 3), I was visiting friends in D.C., so I put up posters around George Washington and Georgetown law schools. The impact was incredible. From a previous weekly high of 4 signups, 21 new users signed up the week of 11/6. I ran the same playbook in a few Long Island law schools the following week, and got 67 new users. This may not seem like many, but to me it was the validation I needed - people were interested and my changes were having an effect.
The best part of these visits was seeing people who initially thought this would be impossible become enthusiastic converts when they saw what the site does. One student pulled up the website, started trying to explain to me that LawStar could never work, then proceeded to try out the internet citation, and her jaw dropped. She sent it to a bunch of her friends right after.
Activation
It’s not enough to get a bunch of signups - users need to successfully find value from the product. I think my activation milestone (the first time a user found success) was copying a citation. The easiest path to a copied citation was driving users to try an automated internet citation. As an illustration of this, consider the chart below - even though there were about 3x more total users than users that used internet citation, users that had done an internet citation were far more likely to have copied 5+ citations.
There were initially a bunch of reasons automated internet citations would break (e.g. they’re trying to cite a PDF, the site doesn’t like bots, non-standard page layouts). I was able to find solutions or build guardrails to these issues so that most new users got a seamless experience upon signup.
The chart below shows copies by citation type, where internet citations are in blue and periodicals are in purple. Immediately after the launch of internet citations on 11/5/23, you can see that internet citations went from being a minority of citations to a vast majority. I changed our homepage defaults to drive people to automated internet citations so that people (especially new users) could feel the magic right away.
Side note - for a while I thought my activation milestone would be downloading the LawStar Chrome Extension. I thought the Chrome Extension would be the easiest way to do a citation.
And because the extension would add buttons to most pages that users did their legal research on, that would organically bring users back into LawStar. The extension supports 4 common research sites, but they never got that much usage. The main thing they’ve been valuable for is reminding users about LawStar. There were a number of cases where users hadn’t come to LawStar in 3 months, and then their revival within LawStar was driven by the Chrome Extension.
Distribution
My strategy to get word about LawStar out there was quite labor-intensive. My most effective technique was to go to schools, put up a lot of posters, and talk to as many students as I could. Thankfully, I look enough like a student that it wasn’t a big deal for me to walk around on those campuses. I thought that after I got some initial cohort of users, I could get them to love the product so much that word would spread in online forums and through word of mouth.
I also experimented with other ways of getting the word out there that never really panned out.
Emailed a number of people at law journals. Sent a couple hundred emails and never got much of a response.
Many law school libraries have pages that point to tools that help with bluebooking, like University of Washington. I got in touch with a few librarians but had trouble pushing them over the edge to actually link to LawStar. I was unhappy about this because this would have dramatically improved my SEO.
Emailed hundreds of legal writing professors. Students learn to bluebook in their legal writing class. No responses.
I got into Law School Reddit and Discord groups and tried to point people toward LawStar when they brought up Bluebooking.
Partnerships with other legal tech firms. Will discuss this more later.
Making Money
There were 3 ways that I thought LawStar could make money:
Subscriptions
Licensing/embedding LawStar’s tech with other legal tech companies for a fee
Charging journals for additional functionality (e.g. collaboration)
I talked to a few journals about what they would pay for and in addition to not really being interested, most journal editors in chief didn’t know where their budget came from, what it was, or how to add a line item to spending. So I dropped that path.
Around the end of December, I had around 300 users. The product wasn’t growing virally, and I didn’t feel comfortable turning on subscriptions yet because I wanted to make sure that I was getting feedback on anything new that I was shipping.
At this point, I thought an acquisition might be the right path come April…and keeping the retained user number higher would look better for future acquirers.
That left trying to monetize through the legal tech companies. I had a 2-pronged approach in these conversations.
They could pay to embed LawStar functionality into their own application in either of the following ways:
I could provide an API for them to build their own experience with.
They could load a snippet of LawStar-operated JavaScript that would modify their UI to make the experience they wanted.
They could not pay and just automatically load the LawStar Chrome Extension plugin code on their page. This would add a “Cite with LawStar” button that would drive people into LawStar to get the citation. This would create a flywheel of people finding LawStar and I could experiment more with monetization of my own subscribers at that point.
Relationship Building
To reach these legal tech companies, I started cold emailing and LinkedIn Messaging people that seemed active in the legal tech space, and tapped my network for introductions to people that knew legal tech. One friend introduced me to the folks at Stanford Codex, where I gave a presentation on LawStar that generated some interesting opportunities. The biggest and most surprising stroke of good fortune, though, was when someone very prominent in the industry, Steve Errick, answered my cold email and proceeded to tell me who to reach out to at almost every major legal tech company with free reign to drop his name. I used to call Steve my fairy godfather of legal tech because his name was like magic - I would instantly get a meeting based on his reference. He also got me a ticket to a major legal tech conference where I was able to talk to a number of other builders that could be possible collaborators.
Once I was in the door, I had conversations with all types of companies about ways to collaborate.
Legal Research Platforms
I had pretty extensive conversations with most of the big players. The pitch I was giving was that many of their citations were wrong, and for some types of documents, 100% of the citations were wrong or missing. Before each conversation, I made sure I had a demo of some cases where their citations failed, and LawStar succeeded within their website so that I could demo the clear improvement.
Ultimately, most of these companies were so focused on AI that they weren’t going to be orchestrating partnership agreements that didn’t further those goals.
Bar Prep Companies
I talked to a few bar prep companies. These companies often sell packages of bar prep courses to schools, so they have the direct line into the schools that might be interested in paying for this. They were interested in the fact that students might come into their site and use their prep courses more if a citation tool was also packaged in the software - the Chrome Extension’s ability to drive people back into LawStar was particularly appealing. The increased utilization would help the bar prep companies in their renewal / sales conversations with the law schools that pay them.
However, the fact that students weren’t paying for LawStar yet led them to be skeptical of the true ROI it would bring to their businesses.
Legal Writing Tools
There are some tools that help with drafting, often tailored at particular legal segments. These companies each want a feature that helps lawyers detect and fix citation issues, as well as create those citations in the first place. I talked to a few of these companies - they all had a strategy that they had recently set to try to solve this problem, so if that panned out then they wouldn’t need LawStar’s help on this problem, and if it didn’t they would probably come back in a few months for assistance on the problem.
Finale
I was pretty confident that if I could hang on for another 6-12 months, I could start to monetize my subscribers in a sustainable way. And a few of the conversations with other legal tech companies would’ve probably developed into 1-2 decent-sized contracts. But I had learned enough, and if I was going to eat glass for another year I may as well focus on a bigger prize.
I let some of those companies and a few others know that I was moving on from LawStar and that they should put together their best offer. One of them ended up being eager to incorporate the underlying technology into its offerings and moved pretty quickly to get into exclusivity. After diligence, the papers were signed and LawStar was officially sold.
Learnings
I won’t spend any time on the standards like “spend more time talking to people” or “listen more than you sell.” And while I learned a ton technically, this isn’t a piece focused on software engineering. So I’ll stick with 3 key learnings that I’m taking away from this - YMMV:
It’s ok to not have a job
When I was leaving my role prior to LawStar, I wasn’t ready to leave and then figure out what was next. I felt like I needed to know what I was going to do before I left. I may have found a more interesting project to work on if I hadn’t felt this way.
I recognize that not everyone has this privilege, but if you have the financial stability to not work for a few months between jobs, do it. Quitting and making networking or validating an idea your full time job can significantly level up your next act. I’ve done that after LawStar and am very happy with the results.
Find the connectors
I underestimated how important it was to find the right people in an industry. I talked to probably 50 people in legal tech. About 40 gave no introductions, and 9 gave 1 introduction. But one person helped me get in touch with ~10 high-quality contacts. A lot of the momentum that LawStar developed with time was due to him - and I will forever be grateful for his kindness and openness.
Don’t be a solo founder
I’m lucky to have a loving and supportive girlfriend, and had roommates that were interested in hearing all the LawStar updates. But there is only so much that people want to talk about your project, while it’s something that you want to think about 24/7. I missed having someone to talk through every situation with, to challenge and sharpen my thinking, and to rely on for support. I tried to fill part of the void with writing pieces like this, but it’s not the same.
What’s Next?
I will share more soon, but the tl;dr is I’m working with a kick ass team of 3 on tech that our customers are amazed by and begging to pay us for. If you’re an engineer looking for your next move and want to build and learn with me and an incredible team, please reach out: jamesjosephmackey@gmail.com.