What autonomy needs
What good autonomy feels like
The one complete handover of the quarter, described from the inside, and the specific things it made possible that the other seven did not.
Decisions this page is about, and where the line actually was
Picking the library for the new parser
1 → 4The architect had a standing view
Dropping a requirement nobody had asked about in six weeks
2 → 5A commitment made before you joined
Changing a database column type
1 → 1Small, reversible, tested once before
- 1Decide it alone and say nothing
- 2Decide it alone and mention it afterwards
- 3Check with one colleague first
- 4Check with the person who assigned it
- 5Needs a decision one level up
- 6Cannot be decided here at all
three decisions · not one of them gave you more room than you thought · two gave you less · one you had exactly right · the median gap was three steps on the scale
Three of the nine decisions counted on this site.
Where ownership crosses several roles, teams can view the feature overview to make time and recurring bottlenecks visible enough for a useful operational conversation.
The wider organisational question can also be compared with resources from Free Management Library.
Of eight handovers on the record counted here, one arrived with all four things. Moving the scheduler onto the new queue. It is worth describing in full, because most writing on this subject describes the failures and leaves the success as an absence.
The brief, roughly as it was given:
The scheduler needs to move onto the new queue. Hard constraint: no more than ten minutes of downtime in one window, because of the support contract. This was started last year and abandoned — the branch is still there, it stopped because the retry semantics did not survive the move and nobody had time to fix them properly. The platform lead has a standing view on queue configuration; show them the plan before you build it. The reason it matters now is that the old queue goes out of support in June.
Four sentences after the first. A minute to say.
What that minute bought
- No exploratory period. Work started on day one instead of day eight, because there was nothing to reconstruct.
- No late veto. The platform lead saw the plan in week one, had two objections, both cheap to accommodate at that stage.
- No rediscovery. The retry problem was known from the first hour, and it was the hard part, so it got the time it needed instead of arriving as a surprise in week five.
- Trade-offs resolved correctly. Knowing the deadline was June and attached to end-of-support meant the shortcuts taken were the right shortcuts.
It is the only piece of work on the record that did not come back.
What it did not take away
This is the part that surprises people who equate context with control.
Every interesting decision was still open. How to do the migration, in what order, with what fallback, how to test it, whether to do it in one step or three, what to do about the retry semantics — all of it was decided by the person doing the work, alone, without checking.
The four things were facts about the world, not instructions. Handing over facts does not reduce anybody's autonomy; it is what makes the autonomy exercisable.
The feeling from the inside
Two things are different and both are noticeable within days.
The first is that you stop watching for the thing you have missed. Working from an incomplete brief involves a continuous background process of scanning for the unstated requirement, and it is tiring in a way nobody has a word for and everybody recognises once named. With four facts in hand, that process switches off.
The second is that asking becomes cheap. When you have been given a real brief, a question is obviously about the work. When you have been given nothing, every question carries an implication — that you cannot work it out, or that you are asking to be told what to do — and so the questions do not get asked.
Why one in eight
Not because seven people were careless. The one complete handover came from somebody who had recently been on the receiving end of a bad one, and said so.
That is the ordinary mechanism. People who hand over well are almost always people who have recently been handed something badly and noticed. The knowledge does not travel any other way, which is why the practice does not spread on its own and why writing the four things down is worth doing.
What it is not
Two misreadings of this example are worth heading off, because both lead somewhere unhelpful.
It is not a specification. Nothing in those four sentences says how to do the migration, in what order, with what fallback, or how to handle the retry problem. All of that was left entirely open and decided by the person doing the work. A specification would have been three pages and would have removed the thing autonomy is for.
It is not thoroughness. The brief is shorter than most of the thin ones on the record, because it contains no background, no restatement of the goal and no description of the system. Four facts and a task. The briefs that fail are frequently longer; length and completeness are unrelated, and conflating them is why "write better briefs" produces documents nobody reads.
What it is, precisely, is four facts that the sender had and the receiver could not derive. That is a short and checkable test, and it is the only property of the message that matters.
Everything else about how it was written — the tone, the order, the absence of headings — is incidental. If the same four facts had arrived as four bullet points, or in three separate messages over an afternoon, the result would have been the same.
The one-sentence version
If you hand work over and remember only one of the four, make it the decider — it is the one whose absence costs the most, and the one nobody ever thinks to say. The last part of this site has the full version, which is a paragraph and takes a minute.