In February 2014, a software company with fewer than 50 employees did something that most investors, most boards, and most founders would consider clinical insanity: it killed off every product it made except one, changed its name to match that surviving product, and told the world it was done growing. No new lines. No adjacent markets. No empire. Just a single project management tool called Basecamp, serving roughly three million accounts, run by a company that had been profitable for over a decade and had never taken a dime of venture capital — well, almost never. Jason Fried, the CEO, wrote a blog post explaining the decision. He could have framed it as strategic focus. Instead he framed it as relief. "We've strayed from the path," he wrote. "Now we're going back."
The announcement bewildered the tech industry not because the company was famous — it was, in certain circles, intensely so — but because the logic inverted every incentive that governed the rest of the software business. The prevailing wisdom held that SaaS companies should be multi-product platforms, that growth came from expansion, that employees were an asset to be accumulated, that capital was cheap fuel for an ever-accelerating flywheel. Basecamp — or 37signals, as it was still known at the time — had spent fifteen years arguing the opposite from a position of quiet prosperity: that small was a feature, not a bug; that profitability was a prerequisite, not a destination; that a company could say no to nearly everything and still matter.
What made the argument stick was the receipts. While venture-backed project management competitors burned through hundreds of millions of dollars chasing growth at any cost, Basecamp had been profitable since roughly its second year of existence. It had never laid anyone off. It paid its employees well, gave them four-day summer workweeks, covered their vacations, and still generated enough cash to fund not just itself but the development of a programming framework — Ruby on Rails — that would go on to power the first versions of Twitter, Shopify, GitHub, Airbnb, and over a million other web applications. All from a company that, at its peak, employed around 75 people.
This is the paradox at the center of the Basecamp story: a company that deliberately chose to be small became, through the software it gave away for free and the ideas it gave away even more freely, one of the most influential forces in the modern internet economy. Not the biggest. Not the richest. But arguably the most ideologically potent — a living, profitable, maddening counter-argument to the entire venture-backed growth model. And also, as events in 2021 would prove, a company whose own internal contradictions could blow up with a force that no amount of philosophical coherence could contain.
By the Numbers
37signals / Basecamp
25+Years in business (founded 1999)
~75Employees at peak
$0Venture capital raised (post-Bezos buyback)
$100M+Estimated annual revenue (reported range)
3.5M+Basecamp accounts created (by 2023)
1M+Web apps built on Ruby on Rails
5Books published by founders
~20Employees lost in one week, April 2021
The Web Designer Who Didn't Want to Be Embarrassed
Jason Fried grew up in the Chicago suburbs, the son of a gemologist — a trade that required discerning the real from the imitation, a skill his son would apply to business dogma with gleeful ruthlessness. Fried started 37signals in 1999, at 25, as a web design consultancy. The name was an homage to a list of thirty-seven candidate signals detected in a search for extraterrestrial intelligence — promising transmissions from deep space, none of which turned out to be aliens. The metaphor was accidental but apt: Fried would spend the next quarter-century transmitting signals about work and business that a certain audience found revelatory and a larger one found confounding.
The consultancy was small. Three, four people. They designed websites for clients — nothing exotic, nothing that required venture backing or a Stanford pedigree. But they had a problem that Fried later described with characteristic bluntness: they kept embarrassing themselves. Projects would slip. Deadlines would pass. Clients would ask for status and the answer was a shrug. Email threads multiplied. Spreadsheets diverged. "Born out of desperate necessity to stop embarrassing ourselves in front of our clients," the company's own origin story reads, "the story of how Basecamp was born is likely your story, too."
It was 2003 when Fried decided to solve the problem. Not by buying software — the existing project management tools were enterprise monstrosities designed for companies a hundred times 37signals' size, suffocated by Gantt charts and resource allocation matrices. Instead, he decided to build something. A tool that would let his small team share files, track milestones, and communicate with clients in a single place. Something simple enough that everyone would actually use it.
He needed a programmer. He found one in Copenhagen.
The Programmer Who Raced Cars and Hated Meetings
David Heinemeier Hansson — DHH, as the internet has long known him — was a Danish business student at the Copenhagen Business School when he discovered Fried's writing online. Before that, he'd been writing about video games on the internet since he was a teenager, had done cold-calling at a web design shop ("Hey, sir, would you like a website?" — this was 1997), and had worked briefly for a venture-backed startup during the dot-com bubble that left him with a visceral, permanent distaste for what he saw as performative business theater: the all-hands meetings, the hockey-stick projections, the burn rate as badge of honor.
He was studying Michael Porter's Competitive Strategy when he emailed Fried. He'd been reading 37signals' blog, Signal v. Noise, which was already developing a cult following for its contrarian takes on web design. Fried hired him — part-time, remote, from Denmark, in 2001. The relationship was improbable: a self-taught Chicago designer and a Danish business student who programmed on the side, collaborating across six time zones with no office, no formal process, and no venture capitalist telling them what to build.
What DHH built was not just Basecamp. In the process of building Basecamp, he created a web application framework — a set of programming tools and conventions — that he extracted from the project and released as open-source software in July 2004. He called it Ruby on Rails. The framework was opinionated, elegant, and ferociously productive. It allowed a single developer to build in weeks what had previously taken teams months. Rails was so good at enabling rapid web development that it became, arguably, the single most important open-source project of the Web 2.0 era. Tobias Lütke used it to build Shopify. Jack Dorsey and Ev Williams used it to build the first version of Twitter. Tom Preston-Werner and Chris Wanstrath used it to build GitHub. The framework powered Airbnb, Kickstarter, Hulu, Square, Zendesk, and Twitch in their early, formative iterations.
I worked partly at a web design shop, where I did such fun things as cold calling for a while. Just calling people up. Hey, sir, would you like a website? The wildest thing was that we actually sold some.
The irony was cosmic. Basecamp — the product — would remain a modestly scaled, deliberately small business. Ruby on Rails — the byproduct, the open-source offshoot, the thing DHH gave away for free — would underpin hundreds of billions of dollars in enterprise value across the technology industry. 37signals chose to be small. Its intellectual output was enormous.
The Bezos Investment That Wasn't Really an Investment
In 2006, something unusual happened.
Jeff Bezos — personally, not through Amazon — invested in 37signals. The amount has never been publicly disclosed, but reports have placed it in the low single-digit millions. It was, by any standard, a tiny check for a man building the world's largest online retailer. And it was the only outside capital 37signals would ever accept.
The investment was notable less for the money than for what it signaled. Bezos was not buying a growth-stage SaaS company with a path to IPO. He was buying a stake in a philosophy. Fried has said that Bezos was interested in 37signals precisely because it was profitable, because it was small, and because it was run by people who had strong opinions about how business should work. For Bezos, who famously told Amazon shareholders that he'd rather be misunderstood than follow the crowd, this was recognizable. For 37signals, it was validation without strings. The investment came with no board seat, no operational control, no obligation to grow.
Later, in a move that underscored the company's allergy to external capital, 37signals would buy back Bezos's shares. The precise timing and terms of the buyback remain private, but the message was loud: even the money of the world's richest man was, eventually, a constraint they preferred to live without.
Profitable on Arrival
Basecamp launched publicly in February 2004. It was not the company's first product — 37signals had previously built a few simple web-based tools — but it was the product that changed everything. Within weeks, paying customers started signing up. The pricing was simple: tiered plans based on the number of projects, starting at $24 per month. No per-user charges. No complex enterprise licensing. No sales team.
The growth was organic and immediate. By the end of 2004, Basecamp was generating enough revenue that 37signals could stop doing client work altogether. Think about what that means: a web design consultancy built an internal tool, released it to the public, and within a year was so profitable from subscriptions that the original business — the thing that paid the bills — became unnecessary. The entire company pivoted to being a software product company, funded not by investors but by customers who paid modest monthly fees because the product was useful.
This is the thing that made Basecamp different from nearly every other SaaS story of the era. There was no "pre-revenue" phase. No years of burning cash to acquire users at negative unit economics. No Series A, Series B, Series C death march toward an IPO that might or might not justify the burn. The product was profitable essentially from launch. The business model was a subscription. The marketing was the founders' writing.
Key milestones in 37signals / Basecamp history
1999Jason Fried founds 37signals as a web design consultancy in Chicago.
2001David Heinemeier Hansson hired part-time from Copenhagen; begins building internal project management tool.
2004Basecamp launches publicly (February). Ruby on Rails extracted and released as open source (July).
2005DHH moves to Chicago. 37signals stops client work entirely; becomes a product company.
2006Jeff Bezos invests (personal capital, undisclosed amount). Additional products launched: Campfire (group chat), Highrise (
CRM).
2010REWORK published; becomes a New York Times bestseller.
2014Company renames itself Basecamp; kills all products except the flagship. Highrise, Campfire, Backpack retired from active development.
The Publishing Company That Happens to Make Software
Most software companies have a marketing department. 37signals had a worldview, and the marketing took care of itself.
Between 2006 and 2018, Fried and DHH published five books:
Getting Real (2006), a manifesto on building web applications with small teams;
REWORK (2010), a New York Times bestseller that distilled the company's contrarian management philosophy into aphoristic essays;
Remote: Office Not Required (2013), which argued for distributed work years before COVID-19 made it a necessity;
It Doesn't Have to Be Crazy at Work (2018), an argument against the cult of overwork; and
Shape Up, a product development methodology book that codified their six-week build cycle. The titles weren't subtle. Neither was the advice.
REWORK sold millions of copies. It was the kind of business book that people either loved or hated — short chapters, declarative sentences, zero nuance. "Planning is guessing." "Meetings are toxic." "Workaholism is stupid." The prose felt like a manifesto written on Post-it notes, and for a generation of small business owners and bootstrapped founders, it was liberation theology: permission to be small, to be profitable, to say no to the growth-or-die imperative that defined Silicon Valley.
We don't like to look too far ahead. We work in six-week cycles. At the end of a cycle, you ship what you've got or you move on.
The books were marketing in the purest sense: they attracted customers who shared the philosophy. If you believed that work should be calm, that meetings were a waste, that 40 hours a week was enough — and you needed a project management tool — you bought Basecamp. The product and the ideology were inseparable. The company's blog, Signal v. Noise, functioned as a perpetual publishing machine, generating organic traffic and ideological allegiance simultaneously. DHH's technical writing on Ruby on Rails brought developers into the ecosystem. Fried's management writing brought business owners. Together, they created a content moat that no amount of paid acquisition could replicate.
The Shape of the Work
The most distinctive operational contribution of 37signals to the broader software industry was not a product but a process. Ryan Singer, who served as the company's head of strategy, codified it in
Shape Up: Stop Running in Circles and Ship Work that Matters, published in 2019. The methodology rejected both the waterfall and
Agile frameworks that dominated the industry.
The core idea was radical in its simplicity. Work was organized in six-week cycles — long enough to build something meaningful, short enough to prevent projects from spiraling into indefinite timelines. Each cycle began with a "pitch" — not a spec, not a user story, but a description of the problem and a rough solution, with explicit acknowledgment of what was out of scope. Designers and programmers worked in small teams of two or three. There were no standups, no daily check-ins, no sprints within the cycle. At the end of six weeks, you shipped what you had or you stopped. No extensions.
Between cycles, the company had a two-week "cooldown" period — time for fixing bugs, exploring ideas, and recovering from the intensity of building. The entire rhythm prevented the chronic problem of most software organizations: the project that was "almost done" for months, consuming resources and morale without producing value.
Fried explained the philosophy with characteristic impatience: plans were guesses, long-term roadmaps were fictions, and the only honest commitment a company could make was to the work directly in front of it. "We work in six-week cycles," he told the REWORK podcast. "At the end of a cycle, you ship what you've got or you move on."
The Shape Up methodology was given away free on the 37signals website — another instance of the company creating enormous value it never directly captured. Product teams at companies far larger than 37signals adopted the framework. It became an alternative gospel in product development circles, particularly among teams that had grown disillusioned with the ritualistic overhead of Scrum.
The Great Renaming, and the Products That Died
By 2014, 37signals had accumulated a portfolio of products: Basecamp (project management), Highrise (CRM), Campfire (group chat), Backpack (information organizer), Ta-da List (to-do lists), Writeboard (collaborative writing), and a few others in various states of maintenance. For a company that preached simplicity, the product sprawl was becoming its own kind of complexity.
Fried and DHH decided to kill it all. Not gradually, not through a slow sunset, but in a single, public declaration. They would focus the entire company on Basecamp — the core product, the thing that generated most of the revenue, the tool that had created the business in the first place. They would rename the company Basecamp. They would stop developing Highrise, Campfire, and everything else, keeping them alive in maintenance mode for existing customers but investing no further energy.
The decision was covered by the Harvard Business Review, which noted its rarity: "Unless you follow tech companies, you might have missed the startling announcement by collaboration and communications software maker 37signals that it has decided to refocus the entire company on a single core product." The HBR framing was characteristically restrained. The tech press was more blunt: this was a company deliberately choosing to be smaller than it could be.
What made the decision unusual was not that a company killed underperforming products — that happens routinely — but that some of these products were performing well enough to sustain their own teams. Highrise had loyal users. Campfire had pioneered real-time group chat before
Slack existed. (The irony of Campfire's premature retirement would sharpen considerably when Slack launched later that year and eventually sold for $27.7 billion.) The founders were not cutting losses. They were cutting possibilities.
HEY, and the Fight with Apple
In 2020, 37signals launched HEY, an email service. If Basecamp was a quiet rebellion against project management bloat, HEY was a loud one against email's capture by advertising surveillance. The product reimagined email from scratch: no tracking pixels, no algorithmic sorting, a screening process that let users decide who could reach their inbox, and a pricing model of $99 per year for personal email.
The product was good. The launch was overshadowed by a fight.
Apple rejected HEY's iOS app update from the App Store, citing a rule that required apps offering subscriptions to use Apple's in-app purchase system — which took a 30% commission. DHH was furious and public about it, turning the dispute into a broader argument about platform power and the "Apple tax." The timing was fortuitous: the European Union was already investigating Apple's App Store practices, and Epic Games was in the early stages of its own legal battle. DHH's complaints became Exhibit A in the emerging case that Apple's platform monopoly harmed developers and consumers.
Apple eventually allowed HEY into the App Store after 37signals added a limited free tier, satisfying the letter of the rules without conceding the principle. But the incident cemented DHH's reputation as the tech industry's most articulate and least diplomatic critic of Big Tech gatekeeping — a role he seemed to relish at 200 miles per hour. (He is, literally, a professional race car driver, having competed in the 24 Hours of Le Mans and the World Endurance Championship. The man's appetite for confrontation was not purely rhetorical.)
The Third That Left
The crisis came fast and from within.
In April 2021, Fried published a blog post announcing several internal policy changes at Basecamp: no more discussion of societal and political issues on company platforms; no more "paternalistic benefits" like wellness and fitness programs; no more committees (including one that had been working on diversity, equity, and inclusion); no more 360 reviews; and a refocusing on the company's core purpose — "we make software." The language was Fried at his most declarative: "It's become too much. It's a major distraction. It saps our energy and redirects our dialogue towards dark places."
The backstory, as reported by Casey Newton and others, was more complicated than the blog post suggested. For years, some Basecamp employees had maintained an internal list of customer names they found funny — many of which were ethnic names of Asian or African descent. When the list came to broader attention within the company in early 2021, it triggered an agonized internal reckoning. A committee was formed to address diversity and inclusion issues. Angry threads proliferated on the company's own internal Basecamp software. Emotions ran hot.
When I talked to employees, what they told me was all of those discussions that they'd been having internally were really about the company itself and how the company could be better and be a more just and equitable organization.
From Fried's perspective, as he later explained, the internal discussions had spiraled beyond productive disagreement into something toxic — "a lot of tension internally, a lot of people distrusting one another." He felt the company was getting away from its founders. The policy changes were, in his telling, an attempt to pull back control and refocus on the work. "I didn't get into this business for this reason," he said on the 21 Hats podcast. "I got into this business to build things, to make things with people who wanted to make things."
The response was seismic. Roughly 20 of the company's 57 employees — about a third — chose to leave, many accepting generous severance packages of up to six months' pay. The departures were public, emotional, and extensively covered. For a company that had literally written the book on calm workplaces —
It Doesn't Have to Be Crazy at Work — the irony was excruciating. The thing that blew up Basecamp wasn't a competitor or a market shift or a failed product. It was an argument about what kind of place it was supposed to be.
We certainly didn't anticipate what was going to happen, which is that about a third of the company decided to leave over this decision. The business didn't suffer at all, which is really interesting.
That last line — "the business didn't suffer at all" — was the most revealing detail. Revenue didn't decline. Customer churn didn't spike. The product continued to work. Within eighteen months, 37signals had hired 45 new people and was larger than it had been before the exodus. The episode demonstrated something uncomfortable about the relationship between culture and business performance: a company could lose a third of its workforce in a week, endure weeks of brutal press coverage, and emerge financially unscathed — perhaps even stronger, if you accepted the founders' argument that the departing employees were a poor cultural fit for the company they wanted to build.
Whether that argument was right or self-serving — or both — remains the central interpretive question of the Basecamp story. The company reverted to the name 37signals shortly after, a symbolic return to origins.
ONCE: Selling Software Like It's 1999
In 2022, 37signals announced something that seemed, even by its standards, willfully anachronistic: a new product line called ONCE. The premise was that instead of paying a monthly subscription for cloud-hosted software, customers would pay a one-time fee and run the software on their own servers. The first ONCE product was Campfire — yes, the same group chat tool they'd killed in 2014, resurrected as a self-hosted, pay-once application.
The ONCE model was a direct challenge to the SaaS business model that 37signals had helped pioneer. In a world where every software company was optimizing for annual recurring revenue, 37signals was selling downloadable software for a flat price. The logic was characteristically contrarian: SaaS subscriptions created vendor lock-in and recurring costs that compounded over time; a one-time purchase gave customers ownership and control; and for 37signals, the economics still worked because the development costs were low (small team, efficient codebase) and there was no customer support infrastructure to scale.
DHH framed it in cultural terms, arguing that the SaaS model had become extractive — that cloud-hosted software had made customers into renters rather than owners, and that the subscription treadmill enriched vendors at the expense of buyers. He linked it to a broader argument about "leaving the cloud" and running software on commodity hardware, a position informed partly by 37signals' own decision to migrate off Amazon Web Services and onto its own servers, reportedly saving millions of dollars annually.
The ONCE line was small. It was not going to dethrone Slack or Microsoft Teams. But it was symbolically potent — a signal (there's that word again) from the company that had helped define web-based SaaS that the model itself might be due for disruption.
The Cathedral and the Bazaar, or: What Ruby on Rails Actually Did
You cannot tell the Basecamp story without reckoning with the asymmetry at its center. The company itself generated somewhere in the range of $100 million in annual revenue — a remarkable achievement for a sub-100-person private company, but not a figure that reshapes industries. Ruby on Rails, the framework DHH extracted from Basecamp's codebase and gave away for free, catalyzed an ecosystem worth orders of magnitude more.
The list of companies that built their initial products on Rails reads like a roster of the modern consumer internet: Shopify (public market cap fluctuating in the range of $100 billion), GitHub (acquired by Microsoft for $7.5 billion in 2018), Airbnb (IPO'd in 2020), Twitter (in its earliest incarnation), Twitch (acquired by Amazon for $970 million), Kickstarter, Hulu, Square, Zendesk, SoundCloud. DHH created, in Rails, a tool that dramatically lowered the cost and time required to build web applications, particularly for startups with small teams and limited capital — exactly the kind of companies that 37signals argued should exist.
The decision to open-source Rails was philosophically consistent but economically irrational by conventional standards. 37signals could have kept the framework proprietary, licensing it for millions. They could have built a Rails-based platform business. Instead, they gave it away, bet that the resulting ecosystem would create demand for their product (Basecamp ran on Rails, proving it at scale), and settled for the reputational capital and developer community that the framework generated.
The bet worked — but not in the way a venture capitalist would have designed it. Rails made 37signals famous among developers, which made Basecamp famous among the kind of technical founders and small teams most likely to buy it. The open-source project was, effectively, the most successful marketing investment in the company's history — and it cost them nothing in direct revenue because it was never meant to generate any.
What Rails also did, less visibly, was establish DHH as a technical authority of unusual stature. His annual RailsConf keynotes became events. His opinions on software architecture, programming languages, and developer culture carried weight far beyond the 37signals orbit. When he criticized Apple's App Store practices, or argued against microservices, or advocated for server-rendered HTML over JavaScript-heavy frameworks, developers listened — not because he ran a large company, but because he had built a tool they relied on.
The Anti-Unicorn Thesis
What does Basecamp prove? This is the question that operators and investors have argued about for twenty years, and the answer depends entirely on what you believe about the purpose of a technology company.
The bull case for the Basecamp model is this: you can build a software business that generates tens of millions (eventually north of $100 million) in annual revenue, remains profitable from nearly its first year, requires no outside capital, never lays anyone off (the 2021 departures were voluntary resignations, not layoffs), provides excellent compensation and working conditions, and does all of this with fewer than 100 employees. The company is not just viable; it is durable. It has survived the dot-com bust, the 2008 financial crisis, the rise of Slack, the COVID-19 pandemic, and its own internal meltdown without existential threat. If the measure of a business is longevity, profitability, and independence, Basecamp is one of the most successful software companies ever built.
The bear case — and it is a real one — is that Basecamp left an astonishing amount of value on the table. The company invented the productivity software category that Asana, Monday.com, Notion, ClickUp, and others would colonize with venture-backed blitzscaling. Asana went public in 2020 with a market cap exceeding $5 billion. Monday.com went public in 2021 at a $6.8 billion valuation. These companies offered products that were, in many cases, more complex and more expensive than Basecamp — but they also captured markets that 37signals had no interest in entering: enterprise sales, platform ecosystems, AI integrations, developer APIs.
Did Basecamp choose not to compete for those markets, or did it lack the ambition? The answer from Fried and DHH has always been emphatic: it was a choice, and a proud one. They didn't want to run a company with thousands of employees. They didn't want to answer to public market investors. They didn't want to spend their lives in board meetings and earnings calls. They wanted to build software, write books, and live well. The company was a vehicle for the life they wanted, not the other way around.
Whether you find that admirable or infuriating says more about you than about the business.
The Calm Company That Wasn't Always Calm
The deepest tension in the Basecamp story is not strategic but philosophical: the gap between the company's public identity and its private reality. For years, 37signals projected an image of serenity — calm work, reasonable hours, thoughtful management, no drama. The books reinforced it. The blog posts reinforced it. The conference talks reinforced it. Fried literally wrote a book called It Doesn't Have to Be Crazy at Work.
Then, in 2021, it got crazy at work. The funny-names list, the diversity committee, the blog post edict, the mass resignations — all of it happened inside a company that had built its brand on being the antithesis of workplace dysfunction. The episode didn't invalidate the philosophy. But it revealed something the books never addressed: that a calm workplace requires either genuine consensus or authoritarian enforcement, and when the consensus breaks, the authoritarianism becomes visible.
Fried's post-crisis narrative was unapologetic. "I'm doing great. The business is doing great," he told the 21 Hats podcast a year later. "Losing a third of your employees in one fell swoop is difficult. But we've recovered. We've since hired 45 new people. We're actually bigger than we've ever been." The recovery was real. The lesson was ambiguous: either the company had pruned dead wood and emerged healthier, or it had traumatized loyal employees and replaced them with people who knew better than to dissent.
The truth is probably both, and the ratio is unknowable from the outside.
Still Transmitting
By the mid-2020s, 37signals — the name restored, the product portfolio expanded again with HEY and the ONCE line — occupied a peculiar position in the technology landscape. It was not a startup. It was not an enterprise. It was not a platform. It was a profitable, private, opinionated software company run by two co-founders who had been working together for over twenty years, who had written their own rules, who had been proven right about remote work and wrong about nothing they'd admit to, and who continued to transmit their signal — about calm work, about small teams, about profitability over growth, about owning your tools instead of renting them — to an audience that had never been larger.
DHH, in a 2026 interview, argued that AI coding tools could not yet match even the most junior human programmers — a contrarian position in an industry racing to automate everything. The stance was characteristically confident, characteristically early, and characteristically impossible to dismiss outright because the man saying it had, after all, built the framework that the industry ran on.
The company still had no plans to go public. Still had no outside investors. Still employed fewer than 100 people. Still sold software that cost less per month than a single lunch in the San Francisco neighborhoods where its competitors raised their venture rounds.
Thirty-seven signals, detected in a search for extraterrestrial intelligence. None of them turned out to be aliens. But someone is still listening.
