Reading Notes

Jason Fried on 37signals, Bootstrapping, and the Shape Up Methodology

Source: Jason Fried on 37signals, Bootstrapping, and the Shape Up Methodology

Notes — Jason Fried on 37signals, Bootstrapping, and the Shape Up Methodology

Four questions [Adler frame]

Q1 — What is it about? The case for a deliberately small, profitable, independent software company — and the operating practices that follow from it. Fried, co-founder and CEO of 37signals (Basecamp, HEY), argues that bootstrapping, small teams, fixed appetites, gut-driven decisions, and refusing to plan far ahead are not constraints to tolerate but the source of the company’s quality and longevity. Threads: why most businesses are not venture-scale; the Shape Up methodology; ‘stay up, not startup’; the ONCE pay-once model; and independence as the root of everything.

Q2 — How is it argued? From 24 years of operating one company against the grain. Hard contrasts carry the argument: ~75 employees and double-digit-million profits versus competitors with 1,000–3,000 staff and thin or negative margins; one product, one price, two-person teams. Where he generalises he hedges honestly — the gut-driven style is ‘easy for me to say, I own the place’; Shape Up is hard to adopt and often fails when forced onto critical work.

Q3 — Is it true? The economic claims about 37signals are first-hand and concrete. The broader prescription is offered as ‘an alternative, not advice to be like me’ — Fried repeatedly concedes survivorship and luck (‘we’re lucky we’ve made it’), which keeps it honest. [?] The ‘venture-backed software has thin margins because it is sloppy’ claim is asserted from observation, not data, and lumps together very different companies. The gut-over-data argument is philosophically strong but, by his own admission, the next book has not yet been written or tested.

Q4 — What of it? Several directly portable practices: appetites instead of estimates (fix the time, vary the scope); two-person feature teams; six-week cycles with a two-week cool-down; adopt new methods on low-criticality work first; stay as small as you can for as long as you can and do the work yourself before hiring; negative visualisation (‘what’s the worst that can happen?’) before a risky bet. The deeper transferable stance: independence buys optionality, and obligations far in the future are dangerous.

Glossary

Stay up (not startup) — Fried’s reframing. Starting a business is easy (‘literally tomorrow’); staying in business for years through plateaus and wavy profits is the hard, uncelebrated skill, and it requires liking the work itself rather than the high of growth.

Appetite (vs estimate) — what you are willing to spend on a feature (capped at six weeks), not how long you think it will take. Estimates expand because work fills the time available; appetites fix the budget and force scope to fit.

Six-week cycle — the maximum duration of any feature, built by two people (one designer, one programmer). ‘Cycle’, not ‘sprint’ — cycles recur like seasons; sprints exhaust and cannot be repeated back to back.

Cool-down — the two weeks after a cycle when teams self-direct: tidy shipped work, fix small bugs, scratch itches, while a few people shape the next cycle’s pitches.

Hill chart — Basecamp’s visualisation of a task as a hill. The left slope is the problem-solving phase (‘still figuring out how’); the top is the turn; the right slope is pure execution. Work still on the left side at the time limit almost certainly dies.

Shaping / trading concessions — designing the work at the right level of abstraction before a cycle (not a spec or tickets), then letting the team define its own tasks and negotiate scope trade-offs as they go. See Shape Up.

ONCE — 37signals’ pay-once, self-hosted, source-available product line: download it, install it on your own server, pay once, no subscription. First product: a relaunched Campfire. See ONCE Model.

Underdo your competition — Fried’s positioning maxim: don’t outdo (expensive, defensive); deliberately do less, with more simplicity and clarity.

Independence (principle 01) — at 37signals.com, the first of 37 stated ideas. No investors, board, or obligations means no external timeframe and no permission needed; every other freedom compounds from it.

Negative visualisation — the Stoic move (Fried credits David Heinemeier Hansson): role-play the worst outcome and make peace with it before committing, so you can move forward unafraid.

Fortune 5 million — Fried’s market: the millions of small businesses, not the Fortune 500. No sales team; self-service; one price.

Key frameworks

Bootstrapping as practice at making money [§ Bootstrapping vs venture]

VC-backed companies practise spending money; bootstrapped ones practise making it — the actual skill of running a sustainable business. Software should carry 80–90% margins, yet Silicon Valley has ‘found a way to make the most profitable style of business the least profitable’ through headcount and customer-acquisition spend. Raise money only when the business genuinely needs capital (factories, hardware, ovens); for software, money buys two things — people and marketing — and mostly buys sloppiness. The guitar analogy: you only get good at making money by practising it; a company pumped full of cash cannot suddenly ‘turn on the profit spigot’. The RXBAR story (‘go sell a thousand bars and come back’) is the antidote.

Small teams and constraints [§ Staying small]

37signals: ~75 people, ~100,000+ customers, one code base, one product per line, two prices on Basecamp, no salespeople, no enterprise tier (max anyone can pay for Basecamp is $299/month, unlimited users). ‘If you gave me 1,000 people I wouldn’t know what to do with them — we’d fall apart.’ Constraints, simplicity, and small teams keep the company honest and the work good. Small is a destination, not a stepping stone.

Shape Up [§ Ways of working]

The 37signals methodology (free book at basecamp.com/shapeup), built on appetites, the six-week cycle, two-person teams, hill charts, shaping, and trading concessions. Work still ‘on the left side of the hill’ at the deadline dies, which prevents the morale-destroying never-ending project. Adoption advice: don’t change a moving company on a dime (momentum is mass × direction); try Shape Up on a low-criticality project first, expect to be bad at it at first, and take whatever parts work — even just six-week cycles or appetites. Full synthesis on Shape Up.

Gut and instinct as a decision tool [§ Decision-making]

Every decision is ultimately a judgment call; unless a machine decides, a human brings thousands of known and unknown experiences to bear. Companies hire executives for ‘judgment and experience’, not spreadsheet-reading. 37signals has no OKRs, KPIs, revenue or growth targets; the only year-end test is ‘did we make more than we spent?’. The operative question is ‘how does this feel?’ / ‘what do you think?’, never ‘how certain are you?’. He wants to write a book celebrating gut and intuition in business; Charlie Munger’s advice when Fried asked about doing something he knew nothing about — ‘trust your instinct’ — is a seed.

Hiring for gut [§ Hiring]

Two signature moves. A paid one-week project for the final five candidates, critiqued together; the key probe — ‘if you had a few more days, what would you do with this?’ — reveals whether the candidate’s gut is full of ideas and whether they will lean into them. And ‘whose work do you admire, and why?’ — chasing the why behind prepared answers to find the candidate’s real lens. He wants clear, well-executed work plus the ability to riff on the spot — not idea-machines who cannot execute, nor executors with no instinct.

Independence and not planning long-term [§ Independence]

Independence is principle 01: no investors means no external timeframe, no required outcome, no permission. 37signals plans six weeks at a time — no one-year, six-month, or quarter plan. ‘I don’t plan long-term because I want to do what I think, not what I thought.’ The further out you plan, the less you know, and obligations in the future trap people into doing things they no longer want to do. Staying close to now is the healthiest, most honest way to run a business. Lenny ties this to the Default Path and Pathless Path; Fried’s life and career were likewise unplanned (‘I’m on a path, I can’t see the path’).

Work isn’t war [§ Words shape work]

Business borrows war metaphors — conquer the market, capture mindshare, target customers, destroy the competition, make a killing — and the words set a depressing frame. Fried prefers additive language: not ‘beat them’, but ‘let’s make something great’. Stop modelling a B2B accounting business on the Navy SEALs. Words establish a way of thinking; a spirit of creation beats a spirit of destruction.

The ONCE model [§ ONCE]

A bet that the SaaS pendulum will swing back amid subscription fatigue: you keep paying the same monthly amount for largely the same software you had four years ago. ONCE products are bought once, downloaded, self-hosted, and shipped with their source code (modifiable, not resellable). The play: find commodity categories still charging luxury SaaS prices, build a simple, high-quality 80/20 version in ~three months with two people, and price it as a no-brainer (under $1,000). Zero marginal hosting cost to 37signals. First product: a relaunched Campfire (a 2006 group chat, pre-Slack). Full treatment on ONCE Model.

Connections