In this article
Before you start
This page answers one question: how do you tell a disagree and commit story well? For the wider topic, see the behavioral interview guide.
Why interviewers ask this
A disagree and commit question separates candidates who can hold a view under pressure from candidates who can also let go of it. Amazon publishes the principle as the 13th of its 16 leadership principles, so a candidate preparing for a loop that borrows Amazon's principles should expect to need a story for it. The guide on Amazon's principles shows where it sits among the others.
A weak answer is a story where you were right and everyone eventually agreed. Another weak answer is a story where you stayed silent and went along. A strong answer shows a real disagreement, the reasons you gave, the decision going the other way, and your full effort afterwards. One coaching guide frames the two halves this way: what you do after you lose the argument, "commit fully, or quietly keep fighting". That is coaching opinion, not Amazon's scoring.
Disagree and commit from first principles
- Consensus trap
- A consensus trap is a situation in which a group will not act until everyone agrees, so a lack of agreement turns into inaction; disagree and commit is a way of avoiding it.
In plain words
Picture friends choosing a film. One wants a comedy, the rest want a thriller. The comedy fan says so once, gives a reason, then watches the thriller and enjoys it. Nobody sulks, and the group did not wait all night. The analogy breaks in one place. At work a decision has costs, an owner and a record, so your reasons matter more.
Disagree and commit has two halves: challenge the decision with reasons, then commit wholly once it is made. Amazon's wording says leaders are "obligated to respectfully challenge decisions when they disagree", and "once a decision is determined, they commit wholly." Silence is not the principle. Half-effort is not the principle either. The Wikipedia entry gives the plain definition used above.
Step 1: Disagree early, with reasons
The first step is to say you disagree, in private or in the planning meeting, before the decision is made. Give your reasons and your evidence once. Amazon's text says to do this "even when doing so is uncomfortable or exhausting".
Input: a decision you think is wrong. Output: your view on the record, backed by facts. In a story, name the data, the risk or the customer effect that worried you. "I had a bad feeling" is not a reason. Respect for the other side helps: an Amazon Web Services strategy post advises disagreeing respectfully and thinking about the other person's responsibilities. You now have a view on the record; the next step is what you ask for.
Step 2: Let the decision owner decide
The second step is to hand the choice to the decision owner and offer an alternative, not a veto. State your view clearly, then propose a version that meets their need. The same Amazon Web Services post suggests proposing an alternative decision framework.
Bezos's 2016 letter gives a sentence for the moment when no one agrees: "will you gamble with me on it? Disagree and commit?". The letter ties this to speed: it says most decisions should probably be made with around 70% of the information you wish you had, and that waiting for 90% usually means being slow. Input: your view plus an alternative. Output: a decision, even if it is not yours.
Step 3: Commit wholly once the decision is made
The third step is the one the story is judged on in practice: you commit wholly once the decision is made. A strong story shows what you did after the decision went against you. Commit means you act as if the plan were yours. You do the work, you answer questions about it without hedging, and you do not reopen the argument in side conversations.
Bezos describes the genuine version: a candid view, a chance for the team to weigh it, and a quick, sincere commitment. His letter says it is not a quiet decision that the matter is not worth chasing (2016 letter, PDF copy). He also applies it to himself: "I disagree and commit all the time." Input: the decision. Output: visible, wholehearted effort.
Step 4: Close the loop with the result
The fourth step is to report what happened, honestly, whether you were right or wrong. Share the measured result with the people who decided. If your concern turned out real, say what you fixed. If it turned out unfounded, say so plainly.
This step turns a story about an argument into a story about outcomes. It also protects the team: a risk you raised is tracked, not forgotten. Bezos pairs the phrase with a related rule: recognize true misalignment early and escalate it immediately. Escalation is for real misalignment, not for a decision you merely dislike. The answer below shows all four steps in one story.
Step-by-step description
- Say it once, early, with reasons.
- The decision owner makes the call.
- Act as if the plan were yours.
- Report the measured result honestly.
- Disagree and commit has two halves: challenge the decision with reasons, then commit wholly once it is made.
- A strong story shows what you did after the decision went against you.
Knowing it is half the job. Saying it is the other half.
Interview Clozo is an AI interview copilot for real interview rounds: it hears the question and puts a strong answer on your screen, and you say it in your own words. See the features or how it works.
Start Free Trial →The Staff-level answer
Tell me about a time you disagreed with a decision and then committed to it.
On a payments launch, I wanted a three-week delay to add retry handling for failed checkouts. I made my case once, with the failure data, and offered an alternative: launch to a small share of traffic first. The product owner chose a full launch on schedule. I said I would back the decision fully, and I did: I built the monitoring, took the first on-call shift and did not reopen the argument. Failures started near four percent. I owned the retry fix in week two, and the rate fell to one percent. In the retro I reported the numbers plainly: my concern was real, but the three-week delay was not needed, because the week-two fix closed it.
This is an illustrative example with invented numbers; replace them with your own story. The answer is about 117 words, roughly 54 seconds at an illustrative pace of 130 words a minute.1 What makes it Staff-level is the trade-off it names (a launch date against a known risk), the ownership of the flagged risk after losing, and an honest close.
The numbers behind the story: in the illustrative example the failure rate fell from 4.0% to 1.0% after the retry fix, a 75.0% drop.1
Text description of the diagram
- In the illustrative example the failure rate fell from 4.0% to 1.0% after the retry fix.
- The illustrative failure rate dropped 75.0% after the retry fix.
A second illustrative example, outside engineering: a sales operations analyst wanted to keep a regional pricing sheet separate, but the partner finance team chose one shared sheet. The analyst gave two reasons once, in writing: regional discounts differ, and one sheet would hide those differences. After the decision, the analyst moved the regional data into the shared sheet, added a tab that kept the discounts visible and trained the regional managers. At the quarter review the analyst reported that the shared sheet worked, and that one discount still needed a manual note.
Two weak versions, for contrast:
- "I told my manager he was wrong, and eventually he agreed with me." (weak) The story only works if you win, and it shows no commitment.
- "I disagreed, but it was not my call, so I stayed quiet." (weak) There was no challenge and no reasons, so there was no backbone to show.
The rubric
| Dimension | Weak answer | Strong answer | Staff answer |
|---|---|---|---|
| Challenge | Complained, or stayed silent | Stated the disagreement to the right person | Raised it early, once, in the right forum |
| Reasons | A feeling or a preference | Specific data or customer effect | Data plus an alternative that met the owner's goal |
| After the decision | Grumbled or kept lobbying | Accepted and did the work | Took on the risk raised and removed it |
| Result | No outcome stated | A result stated | A measured result, including if you were wrong |
| Communication | Rambling, blames the owner | Clear structure, neutral tone | Names the trade-off in one sentence |
For the structure, see how to answer interview questions; for the story frame, the STAR method page.
Three variants of this question
Tell me about a time you disagreed with your manager.
Choose a story where you gave reasons in the right setting and the decision stayed with your manager. End with how you supported the outcome. The interpersonal communication page covers tone.
Tell me about a time you committed to a decision you did not agree with.
Lead with the commitment: what you did, week by week, to make the decision succeed. Mention your disagreement briefly and spend most of the time on your effort and the result.
Tell me about a time you disagreed with a team you led.
Use a story where you were the decision owner and the team challenged you. Bezos says in the 2016 letter that the principle runs both ways, so the boss should do it too. Show that you listened, decided and thanked the dissenters.
Common mistakes
- "I disagreed, then I went along to keep the peace." (wrong) The principle says leaders "do not compromise for the sake of social cohesion". (right) "I stated my case with data, then committed once the decision was made."
- "We split the difference." (wrong) Compromise is a different story. (right) "The owner chose their option, and I carried it out."
- "I committed, but I kept raising it in meetings." (wrong) That is not committing wholly. (right) "I put my concern in writing once, then worked the plan."
- "I was right in the end." (wrong) A win-only story hides the second half. (right) "Here is what I did after we decided, and what the numbers showed."
Follow-up questions
What if you were right and the project failed?
Say what you did when the risk you raised appeared: the fix, the communication, and what you changed about how you raise concerns next time. Avoid "I told you so" and show the work after the failure.
When should you escalate instead of commit?
Escalate when the misalignment is real and unresolved, and do it early; safety, legal or customer-harm issues are examples of the kind of case that fits, and those examples are ours, not Bezos's. His letter says to recognize true misalignment early and escalate it immediately.
Is disagree and commit the same as compromising?
No. Compromise blends two positions; disagree and commit picks one and has everyone carry it out. Amazon's text warns against compromise for the sake of social cohesion.
Who invented disagree and commit?
The origin is disputed. The Wikipedia entry reports Scott McNealy using a version between 1983 and 1991 and an attribution to Andrew Grove at Intel. Amazon added it as a leadership principle in 2010–2011, and Bezos discussed it in his 2016 letter.
Does a disagree and commit story have to be about a boss?
No. It can be a peer, a partner team or a team you led. What matters is a real disagreement, your reasons, and full effort afterwards. For other prompts, see common interview questions and the greatest weakness answer.
- Disagree and commit has two halves: challenge with reasons, then commit wholly.
- Give your reasons once, early, to the right person, with an alternative.
- After losing, do the work, drop the argument and own the risk you raised.
- Close the story with a measured result, even if it shows you were wrong.
- Escalate only true misalignment, early; a decision you dislike is not one.
Sources
- Amazon Leadership Principles — Amazon (aboutamazon.com), 26 Oct 2026
- 2016 Letter to Shareholders — Amazon (Jeff Bezos), 17 Apr 2017
- 2016 Letter to Amazon Shareholders (PDF copy) — University of Washington site (copy of the Amazon letter), 17 Apr 2017
- Guts, Part Three: Having Backbone, Disagreeing and Committing — Amazon Web Services Enterprise Strategy Blog, 26 Oct 2026
- Disagree and commit — Wikipedia (pointer for origin and definition), 26 Oct 2026
- Have Backbone; Disagree and Commit: Amazon Interview Guide — techinterview.coach (coaching site, secondary), 26 Oct 2026
