Disagree and Commit
Disagree and Commit is Amazon's leadership principle for resolving the consensus paradox: how do you move fast without forcing everyone to agree? The answer: surface disagreement loudly BEFORE the decision, then commit fully AFTER, even if you lost the argument. Bezos formalized it in his 2016 shareholder letter when he greenlit an Amazon Studios pilot the entire team thought was a mistake. He literally wrote: 'I disagree and commit and hope it becomes the most watched thing we've ever made.' The principle works because it separates two things organizations usually conflate — the right to be heard and the right to win the argument. You always get the first; you don't always get the second.
The Trap
Two failure modes dominate. (1) Skipping the disagreement: leaders who don't surface dissent get sandbagging — the team executes halfheartedly because they never got their objection on record. (2) Skipping the commit: senior people who lose the argument continue to relitigate it in 1:1s, slack DMs, and side conversations. This is corrosive. Within 6 months everyone learns that decisions aren't actually decisions — they're 'opening positions' subject to constant renegotiation. Both failure modes break trust faster than any other leadership behavior.
What to Do
Operationalize both halves: (1) Before the decision: explicitly invite dissent with 'I want your strongest objection on the table now — what am I missing?' Document the dissent in the decision memo. (2) When making the call: state explicitly 'I'm choosing X. I heard the objections to Y. Here's why I'm going forward anyway.' (3) After the decision: name the commitment expectation: 'I need everyone — including those who disagreed — to commit fully. If you can't, tell me now.' (4) When dissenters fall back into relitigation, name it immediately: 'We made this decision. If you have new information, bring it. Otherwise, commit.'
In Practice
In Bezos's 2016 shareholder letter, he describes greenlighting an Amazon Studios pilot the team thought was a mistake. His exact wording: 'I wrote back: I disagree and commit and hope it becomes the most watched thing we've ever made.' Bezos noted that without this principle, the alternative is 'consensus, which is slow and produces mediocre decisions.' Source: Jeff Bezos, 2016 Letter to Shareholders.
Pro Tips
- 01
Disagree and Commit works UPWARD too. If your boss makes a decision you disagree with, voice it once on record, then commit. The hardest version of this principle is committing fully to a decision you privately think is wrong — but doing it builds the trust that lets you win the next argument.
- 02
Track your 'disagree and commit' calls. After 6 months, audit: how often did the chosen path work? If you 'disagreed and committed' 10 times and the chosen path failed 9 times, you may have been right — and the org may have a decision-quality problem, not a commitment problem.
- 03
Never punish disagreement that was correctly raised pre-decision. The fastest way to kill the principle is to remember who fought you and let it color promotion or assignments. Within 18 months, no one will disagree with you and your decision quality will collapse.
Myth vs Reality
Myth
“Disagree and Commit means giving up your view”
Reality
It means TIMING your view. You disagree fully when the decision is open. You commit fully when the decision is closed. The principle preserves your right to raise NEW information later — it just removes the right to keep relitigating the original argument with the same data.
Myth
“If I commit, I'm responsible for the outcome”
Reality
Accountability for outcomes belongs to the decision-maker, not the implementers. Commit means execute fully — it doesn't mean 'pretend you agreed.' If the decision fails, the leader who made it owns the failure. Healthy cultures separate responsibility for the decision from responsibility for the execution.
Try it
Run the numbers.
Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.
Scenario Challenge
You're a VP. Your CEO has decided to enter a new market against your strenuous objections. You voiced your concerns clearly in the leadership meeting and lost. The CEO has asked you to lead the expansion. You think it will fail.
Industry benchmarks
Is your number good?
Calibrate against real-world tiers. Use these ranges as targets — not absolutes.
Decision Re-Litigation Rate
% of major decisions reopened in hallway conversations without new informationElite
< 10% reopened without new data
Healthy
10-25%
Friction
25-50%
Broken
> 50%
Source: Bain Decision Effectiveness benchmarks + Amazon leadership principle data
Real-world cases
Companies that lived this.
Verified narratives with the numbers that prove (or break) the concept.
Amazon Studios
2016
Bezos greenlit an Amazon Studios pilot the team unanimously thought was a poor bet. Rather than overrule the team or kill the pilot, he wrote his now-famous email: 'I disagree and commit and hope it becomes the most watched thing we've ever made.' He documented the principle in the 2016 shareholder letter as the answer to consensus-driven paralysis. Amazon Studios went on to produce hits including The Marvelous Mrs. Maisel (8 Emmys) and The Boys.
Year Principle Documented
2016 Shareholder Letter
Amazon Studios Subsequent Emmys
27+ since 2016
Prime Video Subscribers (2024)
200M+
Leadership Principles Total
16
Bezos's example matters because it was a CEO disagreeing with the TEAM and committing to the team's view. Most companies use 'disagree and commit' to mean 'employees fall in line.' The principle is bidirectional — it's how leaders commit to teams' decisions they disagree with too.
Netflix
2011-Present
Reed Hastings codified a related principle in 'No Rules Rules': 'Disagree openly. Once a decision is made, support it fully — privately AND publicly.' Hastings calls private complaint after public agreement 'the worst form of disloyalty.' Netflix's 'Farming for Dissent' practice asks every senior leader to actively solicit objections to their proposals before committing. The result: faster decisions and fewer post-decision reversals.
'Farming for Dissent' Practice Codified
2011
Net Subscriber Adds (2023)
29M
Decision Reversal Rate
Low (per culture deck)
Average Decision-to-Execution Time
Industry-low
Netflix's adaptation emphasizes the cultural enforcement: it's not enough to invite dissent — you have to actively farm for it AND publicly punish post-decision sandbagging. The cultural intolerance for hallway politics is what makes the principle real.
Decision scenario
The Dissenter's Dilemma
You're a Chief Product Officer. Your CEO has decided to sunset your most-loved product feature to free up engineering resources for a new bet. You voiced strong objections in the leadership meeting and lost. Your team — who built and loves the feature — is now coming to you upset.
Your Position
Disagreed in meeting
Decision Status
Final
Your Team's Reaction
Upset, demanding context
Your Relationship With CEO
Strong but newly tested
Decision 1
Your team meeting is in 30 minutes. Your eng lead just slacked: 'Is this really happening? I thought you were against it.' You have to choose how to position this with the team.
Tell the team: 'I disagreed with this. The CEO overruled me. I think it's a mistake but we have to execute.'Reveal
Tell the team: 'I had concerns. I shared them. The CEO heard them and decided to move forward. I'm committing to making this work — and I need you to as well. Here's why the CEO's reasoning makes sense even though I'd have weighted it differently.'✓ OptimalReveal
Related concepts
Keep connecting.
The concepts that orbit this one — each one sharpens the others.
Beyond the concept
Turn Disagree and Commit into a live operating decision.
Use this concept as the framing layer, then move into a diagnostic if it maps directly to a current bottleneck.
Typical response time: 24h · No retainer required
Turn Disagree and Commit into a live operating decision.
Use Disagree and Commit as the framing layer, then move into diagnostics or advisory if this maps directly to a current business bottleneck.