Disagree and Commit

There is a failure mode where a decision gets made and nothing moves. The meeting looked settled because nobody objected, and two weeks later the progress check produces "honestly I wasn't sure this was right, so I worked on something else first." Pagedeck hit exactly this after deciding to reduce the free plan's storage limit. Theo said nothing in the room, and Grace read the silence as agreement.

The opposite failure is just as common: teams that keep rescheduling until everyone is convinced. The same item reaches its third meeting, three weeks pass, and the decision is where it started. Past about forty people, any given question touches six or seven of them, and waiting for all six or seven to arrive at the same view means the decision never closes at all.

The two failures look opposite but share a root. They come from treating "raise an objection" and "go along with the outcome" as one bundled act. If disagreeing means you must keep disagreeing, consensus becomes the only available ending. If disagreeing reads as refusing to cooperate, nobody speaks in the first place.

Disagree and commit splits the two apart along the clock. Before the decision, objecting is the obligation. After it, executing is. The principle removes the option of saying "I was against this all along" after the fact — but only by guaranteeing a place to say it beforehand.

It runs on two beats that ask for different things on either side of the decision. Five steps, in order.

1. Open a place for objections — asking "any other views?" once and reading the silence as agreement doesn't count. Go around the room person by person, or have everyone write their concerns into the agenda document beforehand. Quiet people and junior people don't surface objections without one of those mechanisms.
2. Record the objections — write down who objected, to what, and why. Keeping the record means the judgment can be checked against the outcome later, and the person who raised it can see their view wasn't quietly dropped.
3. The named decider decides — not unanimity, not a vote. One person who was identified as the decider before the meeting started. Without that identification this step has nothing to stand on.
4. The people who objected execute too — what's asked for here is contribution, not silence. Doing the minimum assigned share while waiting to be proven right is not commitment. Many teams go further and hand the work to whoever argued against it, on the theory that they see the weak points most clearly.
5. Set the conditions for revisiting — write down, with a number and a date, what would reopen the decision. Something like "if free-to-paid conversion is below 80 percent of the current rate in three weeks, we look again." Without it, the people who objected have no idea how long they are being asked to wait.

Drop either beat and the whole thing collapses. Invoke the phrase without genuinely opening step 1 and it becomes a way to end discussion; skip step 4 and meetings stop being able to conclude anything.

Three situations suit it particularly well.

The first is decisions that are easy to undo. The judgment rule that says reversible decisions should be made quickly rather than debated at length is called a two-way door, and the two principles travel together. Button wording or onboarding order can be reverted next week, so there is nothing to gain from waiting on full agreement. Restructuring pricing takes months to unwind, so objections there get a longer hearing.

The second is decisions with many people attached. Anyone affected by an outcome or able to affect it is a stakeholder, and cutting the free plan's storage touches sales, design, and engineering at once. Sofia has been selling against that limit and Noa has to redraw the signup flow. Rather than waiting for that group to converge, each concern gets heard once and Grace decides.

The third is the retrospective. The record from step 2 gets reopened in the session where the team examines how it works. If objections keep turning out to have been correct on a particular kind of question, the decider's criteria need adjusting. If a whole quarter passes with no objections recorded at all, step 1 has become a formality.

One precondition sits underneath all of this. Step 1 doesn't function without psychological safety — the sense that raising a dissenting view carries no penalty. On a team where the person who objected last quarter is quietly left off the next important project, nobody speaks, and all that remains of the principle is the commit half.

"I'm against this direction. If it goes ahead anyway, I'll own it."

Stating the objection and the commitment in the same breath, which is what the principle looks like when it works.

"If you have concerns, raise them now. After this meeting we're going with this."

Marking that the window for objecting is this one and there won't be another.

"Let's disagree and commit. We'll look at the numbers again in three weeks."

Proposing to move without agreement, with a revisit point fixed in advance.

Pagedeck first ran the full sequence on the free plan change. Grace circulated the agenda document ahead of time and asked everyone to write in their concerns. Sofia wrote that four in-flight contracts would be unsettled; Theo wrote that there was no clean way to apply the change to existing accounts. The meeting took those two first, settled on not applying it retroactively, and Grace made the call.

Sofia remained opposed to the end and still wrote the customer announcement herself the following week. Six weeks later the retrospective confirmed that paid conversion had risen, and her objection was closed with the record left intact.

The most common misreading is that it means doing what leadership decided. Only the second half tends to get quoted, so that's how it sounds. But without the first half the principle doesn't exist. Invoke it without having opened a place to object and the disagreement doesn't disappear — it moves outside the meeting, the decision gets honored on paper, and the resistance shows up instead as execution that keeps slipping.

The second misreading is that a decision, once made, can't be changed. The word commit carries that weight. In practice step 5 is there precisely to fix the conditions for reopening it. Without those conditions, whoever objected has no idea how long they're supposed to hold, and the same item ends up back on the agenda every week anyway.