Autonomy
Concept
How far the people doing the work get to choose the route to a goal, even when the goal itself comes from somewhere else.
Why it exists
At ten people, a founder can look at every decision. Maybe five judgment calls come up in a day, and one person in the company knows enough about all five. Past forty people the arrangement stops holding. Fifty calls a day now come up and the authority to make them still sits in one chair.
Work then fills up with waiting. At Pagedeck, Noa once spent three days trying to change the wording on a button in the checkout screen. Editing the text took ten minutes; asking Grace and getting an answer took three days. Grace had never looked at that screen, so she read the summary Noa had prepared and approved exactly what Noa had suggested. The decision came out the same and three days were gone.
Waiting leaves marks on the person waiting, too. After that happens a few times, people see something worth fixing and leave it alone — not because they doubt they could win the approval, but because the asking costs more than the fixing. That is usually where good suggestions go as a company grows.
Autonomy is the answer aimed at that bottleneck. It doesn't mean making fewer decisions. It means moving the decision to where the information already is. The person who looks at the checkout screen every day, and knows more about it than anyone else in the building, is Noa.
How it works
It works by splitting what from how and handing the second one down. The goal still gets set together or handed down: cut churn two percent this month comes from the leadership, and whether to fix the notifications or fix onboarding to get there is the team's call. Onboarding is the stretch where a new user gets settled into the product.
Handing it down means handing down four things with it. Miss one and what you have is neglect rather than autonomy.
- Goal and deadline — what counts as having done well, and by when. Many companies use a frame like OKR, which writes the objective and the results that confirm it as separate lines.
- Boundaries on decisions — the range someone can settle alone. Spend up to $200 without approval; change the copy on an existing screen at the owner's discretion; escalate anything that touches pricing. The line gets drawn with numbers and examples rather than with adjectives. "Use your judgment" on its own leaves nobody knowing where judgment ends.
- Information — the raw material for the call. Someone who doesn't know what percentage of the quarterly target has come in can hold decision rights and still decide badly. That's why transparency sits in front of autonomy rather than beside it.
- Whether it's reversible — a copy change takes five minutes to undo. A database migration or a price increase does not. Push the first kind toward try-it-and-adjust and the second kind toward decide-it-together.
With those four in place, the manager's job changes shape. It moves from making the calls to maintaining the boundaries and the information behind them. Instead of approvals, the calibration happens in one-on-ones — the regular private conversation between a manager and a team member — where the reasoning behind decisions gets compared.
Where it shows up
It comes up first when deciding how to group teams. A squad is a small team that owns one goal, and a self-organizing team decides for itself how the work gets divided. Both names are designed around autonomy, so when the assumption isn't real only the name survives. Call a group a squad while the lead still makes every decision and all the members learn is that their team has a new label.
The condition that travels with it is psychological safety — feeling that admitting a mistake or disagreeing carries no penalty. Without it, decision rights go unused. If being wrong looks like being alone, people come asking for permission first. A team where authority moved but approval requests didn't is usually in this state.
Once people are spread out, it stops being a choice and becomes a requirement. Under remote work there is no one at the next desk to ask, so the range each person can settle has to be wide enough for work to keep moving.
It also shows up in hiring and retention. A job posting advertising an autonomous environment is really a prompt to ask, during the interview, what a candidate would actually get to decide once inside. An engagement survey usually carries an item asking whether people feel they choose how their own work gets done, and the teams scoring low on it show movement in turnover before anywhere else.
How it sounds at work
"Let's agree on the goal together and leave the method to the team."
The what gets negotiated and the how will not be asked about.
"This one's reversible, so let's just try it and roll it back if it's wrong."
Skip the approval chain here — and treat the hard-to-reverse decisions differently.
"You gave us the autonomy but not the information to decide with."
Decision rights arrived; context did not.
The phrase came up most at Pagedeck right after the teams were split into squads. Mia said she would set the quarterly goal and leave the method to the team, but three weeks in, Theo was still checking on small decisions. From where Theo sat, nobody had said where his half ended. So Mia finally wrote the list out: the team sets the release schedule and the technical choices; the team changes user-facing copy but tells Mia; anything involving price gets decided together. Three lines. The questions dropped off the following week.
What people get wrong
The most common reading is that autonomy means an absence of management. The word sits opposite interference, so it looks that way. In practice, teams with high autonomy have goals and boundaries written down in more detail, not less, because the method can only be delegated once the target is unambiguous. Skip that and "figure it out" sends people three weeks in different directions, and those three weeks get thrown away at the meeting where the work is put back together.
The opposite mistake is assuming that people handed autonomy will naturally do well with it. Someone who receives decision rights without information or boundaries carries every judgment alone and wears down re-checking work out of fear of being wrong. That's the usual route to burnout inside organizations that advertise how autonomous they are. Burnout is the state where sustained overload empties out motivation, and not knowing where the finish line is burns people faster than volume of work does.
Where to go next
- Organizational culture
see which group of conditions autonomy belongs to
- Psychological safety
the state that has to exist before autonomy gets used
Last updated 2026-09-16