Disagree and commit — reduce the decision drag of endless debate

Decisions that are made and executed at 80% confidence are almost always better than decisions endlessly refined but never taken.

Why it works

Decision perfectionism — refusing to commit until all stakeholders are aligned or all information is in — creates decision drag that slows execution, demoralizes teams, and paradoxically often produces worse outcomes than a 80%-confidence decision taken quickly and adjusted based on feedback. The "disagree and commit" norm (popularized independently at Amazon) allows productive disagreement without blocking action.

How to do it

1. Establish a norm: anyone can raise a concern, but once a decision is made, all parties execute it even if they disagreed.
2. Set a decision deadline for significant choices — if no decision by the deadline, the most senior person decides.
3. Debrief decisions explicitly after 30–60 days, not as a retrospective blame session but as a learning loop.
4. Distinguish "this decision is wrong and we should reverse it" (legitimate) from "I don’t like this" (commit and execute).

Evidence

Decision-making research supports the value of committed execution over perpetual deliberation for most reversible decisions; the cost of under-execution of an 80%-right decision typically exceeds the cost of a promptly corrected wrong decision. (mechanistic)

Disagree-and-commit works for reversible, fast-feedback decisions; it is a poor default for irreversible high-stakes decisions where more deliberation genuinely reduces expected cost.

Common mistake

Using "disagree and commit" as a way to silence genuine dissent rather than as a norm that allows productive disagreement to be heard and then set aside — the disagreement must actually be registered, not suppressed.