The most valuable things 37signals ever created were things they made while trying to do something else.
Ruby on Rails, the web framework that runs Shopify, Twitter, GitHub, Airbnb, and about 1.2 million other applications, started as the internal scaffolding DHH built while writing Basecamp. He wasn’t trying to create a framework. He was trying to ship a project management tool for a 12-person web design shop in Chicago. The framework fell out of the process. When he extracted it and put it on the internet for free, it ended up reshaping web development for the next 20 years.
Rework, the book, started as blog posts. Jason Fried had been writing on Signal v. Noise since 1999, mostly short pieces about things that annoyed him about how companies operate. Meetings are a waste. Long-term planning is guessing. Most software has too many features. Nothing about these posts was calculated. Fried wasn’t building a content marketing funnel. He was venting. But the venting attracted an audience, and the audience was exactly the kind of person who’d pay for a project management tool built by someone who hates complexity. Random House turned the blog posts into a book in 2010. It sold over a million copies. Still sells.
Neither Rails nor Rework was planned as a product. Both turned out to be worth more, in terms of influence and downstream revenue, than the thing the company actually sells.
In partnership with T-Mobile
Bring your Business phones. Keep your numbers. Pay $20 a line
Right now, pay just $20 line when you switch 10 or more lines on T-Mobile for Business CoreMobile plan. Bring your device and keep your current numbers.
Plus, get an additional $150/line back when you switch. No limits, so you do the math.
*10 lines for $200/mo; each additional line $20/mo./line with Autopay; plus taxes & $35 device connection charge. $150/line credit via one-time bill credit.
This happens more often than you’d expect.
Amazon built internal infrastructure to handle Christmas traffic. The infrastructure became AWS, which now generates $142B a year and accounts for half of Amazon’s profit. Slack started as an internal chat tool at a game studio called Tiny Speck. The game (Glitch) failed. The chat tool became a $27B acquisition by Salesforce. Google’s 20% time policy (now mostly dead) produced Gmail, Google News, and AdSense. Instagram started as Burbn, a check-in app with too many features. The photo filter was one of those features. They stripped everything else away.
There’s a pattern here, and it’s not “companies should try to make byproducts.” That doesn’t work. You can’t plan a byproduct. The whole point is that it’s unplanned. The pattern is that companies doing hard, real work often generate more value in the margins than in the center, because the center is where they’re focused and the margins are where they’re unselfconscious. The unselfconscious work tends to be more honest, more opinionated, and more useful, precisely because nobody was trying to make it a thing.
37signals is probably the purest example because every major asset the company produced was a byproduct.
The company started in 1999 as a web design consultancy. They built websites for clients. That’s it. Fried, Carlos Segura, Ernest Kim. Agency work. Around 2003 they needed a way to manage their client projects. Nothing on the market worked the way they wanted, which is what you’d expect from a group of opinionated designers. So they built their own tool. DHH, who Fried had found through an email (DHH was in Denmark, 37signals was in Chicago, the whole partnership started with a cold message in 2001), wrote the code.
The tool became Basecamp. The company shipped it in 2004 and gradually stopped taking client work as Basecamp grew. By 2005 they were a software company. By 2014 they renamed the entire company Basecamp because it was the only product that mattered. (They switched back to 37signals in 2022 when they started shipping new products again.)
Basecamp the product is intentionally simple. To-do lists, message boards, file sharing, milestones, a few other things. It doesn’t have Gantt charts, resource allocation tools, time tracking (initially), or any of the features that enterprise project management software competes on. Monday.com has hundreds of features. Asana has hundreds. ClickUp markets itself on having everything.
Basecamp has less. On purpose.
In a normal competitive analysis, this is a weakness. Fewer features = fewer use cases = smaller addressable market. But Fried and DHH turned the limitation into a position. Their argument: software complexity makes work worse, not better. Features accumulate because product teams have to justify their existence by shipping something every sprint. The features compound into bloat. The bloat creates cognitive overhead. The overhead makes people dread opening the tool. By the time your project management software needs a project manager to manage it, something has gone wrong.
This is an opinion. A strong one. Most of the PM software industry disagrees. But the people who agree with it agree with it intensely. And they become Basecamp customers specifically because Basecamp is the only tool built by people who share their view of what work should feel like.
Most SaaS products acquire users through utility. You sign up because the tool does something you need. Basecamp’s acquisition works differently. You sign up because you read Rework and the worldview resonated. Or you read a blog post by Fried about how meetings are toxic and thought “finally, someone who gets it.” Or you watched a conference talk by DHH and decided you wanted to use whatever these guys built.
The product is the secondary artifact. The primary artifact is the philosophy.
In growth terms, this looks weird. Normal SaaS funnels are: awareness (ads, content, SEO) → consideration (free trial, demo) → conversion (paid plan). 37signals’ funnel is: philosophy (blog, books, talks, Rails community) → alignment (reader decides “these are my people”) → trial → conversion. The philosophy layer does what paid acquisition does at other companies, but it’s cheaper (you write the blog posts yourself), more durable (a book stays in print for 15 years), and higher-converting (people who arrive via philosophical alignment churn less because they’re not just buying a tool, they’re joining a tribe).
Fried has said 37signals spends basically nothing on traditional marketing. They don’t buy billboards or run paid social or bid on keywords (except to defend their own brand name, after competitors started buying “Basecamp” on Google). The books, the blog, the podcast, the conference talks, the open-source contributions to Rails, these ARE the marketing. And they cost whatever it costs to write and show up, which for a company with $280M in revenue and ~170 employees is essentially a rounding error.
The Bezos investment in 2006 fits the same logic. Bezos took a minority stake. No board seat. No control. Fried treats this as buying optionality: the investment gave 37signals capital flexibility without selling decision-making power. Every other growth-stage SaaS company takes VC, gets board members, and eventually faces pressure to IPO or sell. 37signals faces none of that. The investment structure is itself an expression of the philosophy. Stay independent. Keep control. Optimize for decades, not quarters.
The question people ask about 37signals is whether the model scales. And the answer is: it scales to exactly the size they want it to be, and no further.
Basecamp probably can’t become a $10B company. The addressable market for “simple project management for people who hate project management” is real but bounded. Monday.com is worth $12B because it went after enterprises with hundreds of features and an aggressive sales org. Basecamp explicitly rejects all three: the enterprise market, the feature arms race, and the growth-at-all-costs playbook.
But what’s interesting is that the model doesn’t need to scale like that to be extraordinary. 37signals has been profitable for 25 consecutive years. Tens of millions in annual profit on roughly $280M in revenue. No debt. No investors to answer to. The founders kept ownership and control. They’ve shipped multiple products (Basecamp, HEY email, the ONCE line of buy-once software, Campfire which they open-sourced in 2025). They’ve written bestselling books. They created a web framework used by millions of developers. They pioneered remote work years before COVID made it mainstream.
By the metrics that actually matter to the people running the company (profit, autonomy, creative freedom, impact relative to headcount), 37signals is one of the most successful software companies ever built. It just doesn’t look like one if you measure success by valuation multiples and headcount growth.
And the cloud repatriation move in 2023 is a perfect distillation of the philosophy. DHH did the math on their AWS bill, concluded it was cheaper to buy and run their own servers, moved off the cloud, and saved $1 million a year. Every other company in tech was moving to the cloud. 37signals moved the other direction because the numbers said so and they had no board members or investors who’d object for optics reasons.
Then they launched ONCE, a line of software products you buy one time and self-host, as a direct refutation of the SaaS model that they helped popularize. The company that built one of the earliest successful SaaS products now argues that SaaS is often a bad deal for customers. They went from “pay us monthly forever” to “pay us once and own it.” That kind of move is only possible when you don’t have a board demanding recurring revenue growth.
There’s a tension in 37signals’ story that Fried and DHH don’t fully resolve, and it’s worth sitting with.
In 2021, the company banned societal and political discussions in internal forums. Fried called them “a major distraction.” A third of the company resigned over it. The departures were public, messy, and got significant press coverage. Critics called the policy authoritarian. Fried didn’t reverse it.
This is the shadow side of strong opinions. The same conviction that makes Rework compelling, that makes Basecamp opinionated, that makes Signal v. Noise worth reading, also produces a company where the founders’ worldview is non-negotiable. When employees’ politics conflicted with the founders’ preference for a politics-free workplace, the founders chose the philosophy and accepted losing 33% of their team.
You could read this as principled. You could read this as rigid. It’s probably both. The same quality that makes 37signals 37signals (founders who believe things strongly and build accordingly) is also the quality that makes it a difficult place to work if you don’t share those beliefs. The tribe that makes the marketing work also makes the culture narrower than most companies would tolerate.
For a company built on the idea that work should be calm, losing a third of your staff in a week isn’t calm. But Fried would probably argue the disruption was temporary and the clarity was permanent. Once you’ve established that the company’s philosophy comes first, you don’t have to relitigate it every time.
The question is whether a company can be built on opinions so strong that they function simultaneously as the product’s greatest asset and its greatest risk. After 25 years and counting, 37signals suggests the answer is yes, as long as you’re willing to accept the limits that come with it. You’ll never be a $10B company. You’ll never please everyone. You’ll lose employees who disagree with you. But you’ll also never have to pretend to believe something you don’t, which is worth more than most founders admit.
Further reading
- Why We Choose Profit at 37signals (37signals): The company’s own manifesto. Zero hours spent on valuation. Simple LLC structure. Why being profitable when the market crashes is the whole point.
- Jason Fried on 20 Years Bootstrapping Basecamp (Practical Founders, 2025): Fried in his own words. The “would I hire them again” question. Why small teams build better products. Why they reject the enterprise market.
- Case Study: 37signals (Library of LLM): Comprehensive overview. Revenue per employee. Shape Up methodology. Rails as a byproduct. Cloud repatriation. The contrarian position on every axis.
That’s all for today. I’m going to go stare at my laptop and pretend I’m thinking strategically. See you Monday.
Leave a Reply