Bootstrapped to $1.8M ARR selling software to small law firms

The lowest point was a stretch in summer 2024. Three things happened in six weeks. Our biggest customer — a 5-attorney firm in Sacramento paying $495/month for an enterprise tier — churned because their managing partner retired and the new partner had a nephew who "did computers." A state bar I'd flown to threw us off the expo floor because we hadn't paid an undisclosed sponsorship fee. And Theo's wife was due with their first kid in August.

I was in a hotel in Phoenix when Theo called and said, very evenly, "I might need to go take a real job for a year. I can't do this and a newborn." I sat on the bed in that hotel and did mental math about whether I could run the company solo while Theo took a sabbatical.

§The Knoxville email

What saved us was an email from a customer in Knoxville: "Just wanted to say my paralegal Diane cried when she saw what your product does. We're recommending it to three other firms in our bar association." Theo and I read that email out loud to each other. He decided to keep going. Diane is still a Casefile user as of April 2026. She is also, as it turned out, the reason we bothered to build a referral mechanic into the product the following year.

In February 2025 we did something I'd been resisting for a year. We doubled our prices.

For two years we'd been priced like a SaaS app — $99 per attorney per month, $39 per support staff. The math worked but slowly. I'd been told by every other vertical SaaS founder I knew that we were underpricing. I didn't want to believe them because the customers were small firms and price felt mean.

We moved to a per-firm model: $399/month for solo and 2-attorney, $799 for 3-5 attorney, $1,499 for 6-10. Existing customers grandfathered indefinitely on Stripe's legacy price IDs. New customers, new prices.

The pipeline didn't collapse. New-customer conversion dropped about 18%. Revenue per new customer roughly tripled. By August 2025 we'd added $640k in new ARR in six months — more than we'd booked in the previous two years combined. As of April 2026 we're at about $1.8M ARR, profitable, and the company is six people: me, Theo, two engineers, and two customer-success folks who used to be paralegals. Stack is still Next.js, Vercel, Postgres, Stripe, Resend, with Linear for issues and PostHog for product analytics.

We still don't take outside money.

§What changed when the buyer changed

The first six months after we re-cut the ICP from "solo attorneys" to "two-to-eight lawyer firms with a paralegal who hates the current PMS," everything moved. Annual contract value tripled from $1,200 to $3,600. CAC payback fell from fourteen months to four. Trial-to-paid lifted from 9% to 31%. The reason: paralegals are the actual end-users, and they have the authority to recommend a switch from Clio, MyCase, or PracticePanther to the managing partner — who almost always defers to them on tooling.

We hired Marisol, a paralegal-turned-implementation-lead, in late 2024. She did every onboarding call for the next nine months and her win rate on "considering Clio vs Casefile" was 78%. She'd been a customer first. That hire mattered more than the three engineers we hired after her combined. We crossed $1.8M ARR in March 2026 with twelve people, no venture funding, and a 92% gross margin. The bootstrapped vertical-SaaS playbook from the embedded-payments fintech story in this cohort maps almost directly onto legal — different industry, identical motion.

1. 1
   "Impossible markets" are usually just unsexy markets that VCs can't model.

   Small law firms aren't impossible. They're just below the AOV that fits a venture portfolio. That's not the same thing.

2. 2
   The state-bar expo circuit is dead-simple distribution if you'll actually do it.

   Twelve weekends a year, a folding banner, and a willingness to talk to people whose first question is "do you take checks."

3. 3
   Per-firm pricing beats per-seat for SMB vertical SaaS.

   Small businesses don't track seats. They track invoices. One predictable invoice converts dramatically better than a per-seat model that scales with their hiring.

4. 4
   Grandfather aggressively.

   Letting existing customers keep their old price is the cheapest insurance against the only real risk of a price change — public anger.

5. 5
   Bootstrapping is a strategy, not a virtue.

   We chose it because it gave us permission to ignore growth math investors would have imposed. If your market wants you to grow at venture speed, raise. If it doesn't, don't pretend it does.

What this looked like in practice.

The thing nobody tells you about *vertical SaaS* is that the first ICP you write down is almost always wrong by one degree. We thought solo attorneys were the buyer; the actual buyer was the paralegal at a slightly larger firm. Re-cutting the ICP took eight months of customer calls.

The thing nobody tells you about *competing with Clio* is that you do not compete on features. Clio has a thousand-person team and twenty years of head start. We compete on "your paralegal will not quit because of this software." That is a different product category.

The thing nobody tells you about *paralegals as champions* is that they are systematically under-served by every other tool. Selling to them is the highest-leverage GTM move in legal tech, and almost nobody does it because paralegals do not show up in pitch decks as "decision-makers."

The thing nobody tells you about *bootstrapping in legal tech* is that the bar association referral networks are slow but compounding. Our second year of growth was 80% word-of-mouth from a referral pattern that only became visible in retrospect.

The thing nobody tells you about *charging $300/month per seat* is that price is a signal of seriousness. We tried $79 first. Nobody bought. We tried $299 and people bought twice as fast — because the higher price implied we would not disappear.