Jason Fried: The Anti-Exit Playbook for $80M Revenue

Most companies treat constraints as failure states they're trying to escape. Fried treats them as strategic choices that create competitive advantage. Basecamp doesn't have features Asana has — and that's exactly why some teams choose Basecamp. The simpler product fits a specific use case better than the loaded product fits any. Picking your constraint is positioning.

Basecamp was built for 37signals' own use, then sold when clients asked "what is this thing?" HEY was built because Fried and DHH were unhappy with email. ONCE was built because they were unhappy with the SaaS subscription model they helped create. Building for yourself produces natural product-market fit — the constraint is your own use case.

Ruby on Rails (2004) and Campfire-as-open-source (2025) are not direct revenue plays — they're talent pipelines and credibility moats. Releasing infrastructure for free generates goodwill that compounds for decades. The Rails community has produced engineers, customers, and a permanent technical reputation that no marketing budget could buy.

Primary investment puts capital into the company — which then has to deploy it, account for it, and generate returns on it. Secondary investment buys existing equity from founders. The company doesn't change financially; the founders take chips off the table without changing operations. This is the structural difference that lets the deal preserve everything else: no capital needs to be deployed, no growth rate needs to be hit, no burn justified.

VC investors require exits because their funds have 7–10 year lifecycles. Bezos isn't a fund — he's a long-horizon individual investor. The deal pays him via annual profit distributions on the LLC, the same way it pays the founders. 18+ years of distributions on a $10M stake have already returned the principal multiple times — and he still holds the equity. The structure aligns infinite-game incentives.

DHH wrote: "The hardest part was actually talking Jeff down on the size of the purchase. Originally he wanted a bigger slice — which would have meant more money for Jason and I, but it would also have meant giving up more of the company." The lesson: when negotiating non-exit equity, the goal isn't to maximize cash today — it's to minimize the equity you give up. Smaller stake at the same valuation means more upside retention.

Almost no one else has structured a deal like this with a comparable investor. Why? Because it requires (a) a profitable business that doesn't need primary capital, (b) an investor willing to take returns via distributions rather than exits, and (c) founders confident enough to refuse standard terms. Most founders fail filter (c) — they accept standard VC terms because they don't know the Bezos pattern is even available.

Each product has its own brand, pricing, and customer audience — but engineering, design system, support staffing, and infrastructure are pooled. This is what lets fewer than 80 people run 4+ products at $80M+ revenue. A conventional structure would have product-specific teams that scale with each product; the holding-company structure scales operations once for all of them.

In 2014, 37signals shut down Highrise, Campfire (as a product), and Backpack to refocus on Basecamp. Most companies treat product closure as failure; 37signals treats it as portfolio discipline. When complexity exceeds value, you concentrate. The 2014 consolidation was what enabled the 2020 expansion (HEY) — focus before you expand again.

The 2022–2024 cloud exit (AWS to owned hardware) is the holding-company logic applied to infrastructure. Owning your own servers is unfashionable but structurally aligned with the no-VC, no-exit model. Projected $10M+ in 5-year savings; permanent insulation from cloud-vendor pricing changes; control over the entire operational stack. Same logic as the Bezos deal: don't accept structures that create future obligations you don't need.

Constraints that come from outside (not enough customers, not enough capital) feel like failure. Constraints chosen on purpose feel like strategy. 37signals chose a constraint — small team, simple products, no outside capital — and treated it as the moat. The constraint preserves optionality forever; capital extracts optionality immediately.

The default assumption is "take VC = lose control." The 37signals counter-example proves there's a third path: secondary minority investment with no board, no liquidation preference, and profit-distribution returns. This requires a profitable business and a long-horizon investor — but if you have both, it's negotiable. Don't accept standard VC terms because you didn't know an alternative existed.

Fried: "Pull out money along the way so you have something to show for your work, because there's a very good chance it isn't going to work." The infinite-game posture isn't "work for free until exit." It's "profit annually, distribute proceeds, and let the equity appreciate as a separate asset." Distributions de-risk the founder. They also align everyone (including outside investors like Bezos) on annual outcomes rather than terminal exits.

Signal v. Noise (1999), REWORK (2010), REMOTE (2013), Shape Up, It Doesn't Have to Be Crazy at Work — 37signals has published more thought leadership than most companies five times its size. Each book and post is also a customer-acquisition vehicle. The teaching IS the marketing. No advertising budget needed because the audience is built through the publishing.

Basecamp, HEY, and ONCE were all built to solve problems 37signals had internally. Selling those tools to others required almost no validation step — they were already validated by their own use. Building for yourself is the cheapest, fastest, most accurate market research method available. It also ensures you have at least one passionate user from day one: yourself.

Jeff Bezos as the minority investor. The 2006 deal terms — secondary purchase, no board, no liquidation preference, distribution returns over 18+ years — required a long-horizon individual willing to take returns through profit distributions rather than an exit. That investor profile is rare and not on the open market. The DHH partnership. David Heinemeier Hansson didn't only co-build Basecamp; he wrote Ruby on Rails inside the company and gave 37signals permanent technical authority no marketing budget could buy. Most founders will not pair with someone who invents an industry-defining framework while building the product. 1999/2003 SaaS founding-era window. Launching a project-management SaaS the same day as Facebook, into a market where subscription pricing was still novel, gave 37signals two decades to compound before category saturation. That window has closed. Profitability before capital. The Bezos terms only worked because the company didn't need primary capital; founders without an existing profitable business cannot negotiate from that position.

But the anti-VC architecture is universal. Pick a constraint and treat it as the moat — the simpler product is a category of one when everyone else is bloating their feature set. If capital becomes available, structure it as secondary minority investment with no board and no liquidation preference rather than accepting standard VC terms. Distribute profits annually rather than working toward a terminal exit; the distribution discipline de-risks the founder and aligns all stakeholders on infinite-game outcomes. And use teaching as the marketing — every book, post, and open-source release compounds for decades. These principles work whether the team is 80 people or 8.