Jason Fried's Appetites: Fixed Time, Variable Scope

Fried is careful to flag that the swap sounds smaller than it is.

An estimate is a prediction about the work. It treats scope as fixed and time as the unknown. Fried's position is that the prediction is almost always wrong. When it is wrong, the only thing that can give is the schedule. That is the mechanism behind projects that keep going.

An appetite inverts which variable can move. The team fixes time up front, based on what the problem is worth to the business. Scope is the unknown, and the team's job becomes finding the version of the feature that fits inside the budget.

Not everything gets the full budget. Fried describes one-week appetites for things that are not that big a deal, where having the feature is nice and not having it is fine. The size of the appetite is itself the prioritization decision, made before any code exists.

An appetite you can extend is just an estimate with better branding. The rule that gives it teeth is that work which does not ship inside the bet does not get an automatic extension. It stops, and it has to re-compete for a future cycle against everything else.

Fried's argument for the circuit breaker is not about efficiency. It is about what open-ended work does to people.

Boring work is survivable when it has a boundary. The same work with no visible end is the thing that makes people want to leave. Fried is explicit that a nine-month project someone is four weeks into is a different psychological object than a four-week project. This holds even when the task is identical.

A two-week cooldown follows the cycle, and he picks the vocabulary with care. He rejects the word sprint, because you cannot go all out again after sprinting. Cooldown is when teams breathe, fix what they broke and decide what to bet on next.

1. Set the appetite before the discussion of scope. Ask how much time this problem is worth, not how long it will take. Six weeks is the ceiling; one or two weeks is common.
2. Shape the work before handing it over. Senior people define the problem and the boundaries. An unshaped problem handed to a team turns the appetite into a guess again.
3. Give it to a small integrated team. One designer and one or two programmers, with full responsibility for finding the solution inside the budget.
4. Cut scope, never extend time. When the work will not fit, the feature gets smaller. That is the whole trade, and it is where most adoptions fail.
5. Let the circuit breaker fire. If it does not ship, it stops. No automatic rollover. The project has to earn a new bet like everything else.
6. Take the two-week cooldown. Not a buffer for spillover work. Time for cleanup, exploration and choosing the next bet.
7. Keep no backlog. Fried does not maintain a growing list. If an idea matters, it will come back on its own. If it does not come back, it was not important.
8. Pilot it somewhere low-stakes. His own recommendation: test the method on a low-criticality project first, because a high-stakes failure turns the team against the process.

It works when scope is negotiable and you control the roadmap. That is why it fits product companies and bootstrapped teams so well. Nobody has promised a specific feature set to a specific customer on a specific date.

It fails when scope is contractual. Enterprise commitments, regulatory deadlines and platform migrations do not let you ship two thirds of the thing. It also fails when leadership treats the appetite as a target rather than a boundary. That recreates the deadline death march Shape Up exists to prevent.

And it fails on adoption more often than on mechanics. Fried is candid that trying to turn a company 180 degrees does not work, because existing momentum is real. Companies that mandate Shape Up organization-wide on a Monday tend to abandon it within two cycles.

Where operators split. Fried's method rests on judgment: someone decides an appetite up front. He defends intuition over data-heavy analysis, on the grounds that people are feeling creatures. Ray Dalio would push the opposite way, weighting that same call by who has a track record and can explain the cause and effect. The split is real and it is about evidence density. Dalio's approach needs a history of comparable decisions. Fried is working in the space where that history does not exist and waiting for it costs more than being wrong.