When everything is a priority, run the capacity line

"It's all P1." Every engineering manager has heard it, most have said it, and the standard diagnosis — our stakeholders can't prioritise — is wrong in a useful way. The stakeholders are behaving correctly. The system they are behaving inside is missing a part.

An ordering costs nothing

Priority is an ordering, and an ordering is free. Calling something P1 displaces nothing: no other item moves, no capacity is consumed, no one is told their thing lost. So of course everything becomes P1 — the label is the only lever a stakeholder has, it costs them nothing to pull, and pulling it harder than the next person is the rational move in a queue where only loudness ranks.

Re-ranking the list does not change this. Neither does a scoring framework — a RICE score is an ordering with decimal places, and it inherits the same free-ness: scoring an item 5.3 displaces nothing either. The room feels rigor and the queue feels nothing.

A constraint is what converts ranking into deciding

Put the same list against a per-team capacity line — each team's honest FTE-months for the quarter, standing costs already subtracted — and walk it down in order, subtracting per craft as you go. Somewhere, usually sooner than anyone likes, a team's line is spent. At that point "everything is P1" collapses on contact with arithmetic: the eleventh P1 has to name which of the ten above it moves, out loud, to a person who was promised it.

That sentence — which of these moves? — is the entire technology. It converts an argument about labels into a decision about trade-offs, and it is checkable by everyone in the room: "these two don't both fit in the backend team's quarter" is arithmetic, not opinion.

What happens to the phrase

Stakeholders do not stop wanting everything — wanting everything is their job. What changes is what the want has to pass through. Three exits, all of them honest: the item fits, and something below it moves down visibly; the item does not fit, and it gets a recorded not-now — parked with a review date or declined with a reason; or the item genuinely cannot wait, and it crosses the line as a recorded exception — why now, what is missing, who accepts the risk — with the displacement it causes made visible in the same breath.

Notice what none of the exits is: a fight about whose thing is "really" P1. The label stops mattering, because the label was never the scarce thing. Capacity was.

The mid-quarter version

The same mechanism handles the urgent ask that arrives in week five wearing a P1 badge. Capacity did not grow because a request arrived — so the ask is an edit to a named commitment, never an addition. "Yes, and it displaces the retention work — shall I tell its stakeholder, or will you?" is a complete answer, and it is remarkable how many P1s discover they can wait until next quarter when the displacement has a name attached.

Why this beats being better at pushing back

The standard advice for priority inflation is interpersonal: hold the line, manage up, learn to say no. It makes the outcome depend on the manager's stamina, and stamina loses to seniority every time it matters. A visible line moves the argument off the person and onto the record — the manager is no longer refusing, the arithmetic is — which is also why an over-committed team is the same conversation in a later, more expensive form: by then the displacement already happened, silently, and someone is about to find out which commitment paid for it.

What DeliverySheet holds is the record that makes the line checkable: every commitment carries the per-team capacity figures the delivering teams gave, the committed-capacity report sums them per team so the walk has honest arithmetic, committing before the evidence is ready requires the recorded exception, and the quarter's exceptions are countable at its end. The line itself is a discipline you run in the planning room — and "everything is a priority" stops being a culture problem the moment it has to fit under a number.