Tell me about a time you failed

Why interviewers ask about failure

Partly because everyone has failed, so the question is a reliable way to see how a candidate behaves when there is nothing flattering to report. Harvard’s career services makes the point directly: everyone has weaknesses, and everyone has made mistakes.

Three qualities show up in the answer. Accountability, which is whether you can describe your own contribution without reaching for circumstances. Awareness, which is whether you understood what went wrong well enough to explain it. And correction, which is whether anything about how you work actually changed.

There is also a filter at work. A candidate who cannot produce any failure either has not reflected on their work or has decided to manage the interviewer rather than answer. Both are more worrying than a mediocre example honestly told.

MIT’s sample questions include the starker pairing, “Define success. Define failure.” If you get that version, your definition of failure is doing the same work as the example would: it tells the interviewer what standard you hold yourself to.

Choosing the right failure

The selection is most of the work. Run your candidates through five tests.

Were the stakes real? A failure with no consequence reads as a dodge. Something that cost time, money, a client’s patience or your own credibility has enough weight to be believable.

Is it finished? Choose something resolved, where you can describe the ending and what you have done since. A live problem puts the interviewer in the middle of your current difficulty.

Was your part real? If your honest account is that someone else caused it, the answer collapses. Pick the one where you can name a decision of yours that contributed.

Is it survivable? Avoid failures that disqualify you from this specific job. A missed deadline is fine almost everywhere; a medication error is a bad choice for a nursing interview, and a data breach is a bad choice in security. MIT’s guidance on weakness questions applies here too: avoid discussing a weakness that is important to the role for which you are applying.

Can you name what changed? One concrete change, such as a check you added, a habit you keep or a conversation you now have earlier, is the part the interviewer remembers.

One more point of calibration. The failure should be about judgement or execution rather than character, which is why Harvard advises focusing on behaviours rather than personality traits. “I misjudged how long the migration would take” is a behaviour you can fix. “I’m careless” is a self-diagnosis that invites doubt about everything else you said.

Failures to keep out of the answer

Anything involving dishonesty. Expenses, falsified figures, covering something up: interviewers cannot hear these as learning stories, however well you tell them.

Anything that endangered someone, in any job where safety is part of the work. Even with a good recovery, it becomes the only thing they remember.

Anything confidential. Naming a client, a patient, a pupil or an unreleased product is itself a failure of judgement happening in front of them. Say the sector and the size, not the name.

Someone else’s failure. Answers that spend ninety seconds on a colleague’s error, with you as the person who noticed, are the most common evasion and the most obvious.

The fake failure. “I failed by caring too much” or “my failure is that I take on too much work” are weaknesses in costume, and a dressed-up answer wastes the credit you would have earned for a real one.

And the catastrophe with no recovery. If the story ends with you being dismissed and nothing else, there is no second half. Choose something where you did something about it, even if that something was asking for help late.

How to structure the answer

Four beats, with the weight at the end.

What you were trying to do, in one or two sentences. Give just enough context for the failure to make sense: the goal, the deadline, who was depending on it.

What went wrong and what your part was. Say it early and plainly, in the first person. “I underestimated the testing and did not ask for a second pair of eyes” is the sort of sentence that buys you credibility for the rest of the answer. Mention external factors once, briefly, then return to your own decisions.

What you did immediately. This beat is often missed and it matters a great deal: who you told, how fast, what you offered, how you limited the damage. A candidate who told their manager the same morning is describing a professional habit, not a confession.

What changed since. End here, and make it specific. MIT’s STAR guidance notes that the Result section can emphasise what you learned from the experience or your key takeaways, which is exactly how a failure answer should close. Then, if you have one, add the proof: a later situation where the new habit held.

Spoken, that runs about ninety seconds. Keep the first two beats tight, because the last two are what the interviewer is scoring.

Sample answers

Missed deadline

My clearest failure was missing a client reporting deadline that I had promised twice.

I was building a quarterly analysis for an insurance client, due on a Friday. I knew by the Tuesday that the data from their broker file was in a different format than expected, and I decided to push through rather than flag it, because I thought I could still finish.

I could not. I told the account manager on Friday morning, which left her no room to manage the client, and she had to apologise for me. I worked the weekend and delivered on the Monday with a written note of what had been simplified.

What changed is that I now treat any surprise in the inputs as a reporting event, not a problem to absorb quietly. On the next quarter’s run I flagged a similar file issue on day one, and we renegotiated the date before anyone was promised anything.

A hire that did not work

The failure I think about most is a hire I made badly.

I needed a second analyst quickly and interviewed a candidate who was strong technically and vague about why he wanted the role. I wanted the vacancy closed, so I read the vagueness as modesty and made the offer.

He left after four months, and in the exit conversation it was obvious he had never wanted the job’s reporting side, which was most of it. The team lost a quarter and I had to restart the search.

Since then I ask every candidate to describe the least interesting part of the role in their own words, and I do not make an offer when I cannot explain the person’s motivation to someone else. The two hires I have made since have both stayed.

Frontline or service failure

Early in my time on a support desk I promised a customer a refund I was not authorised to give.

She had been passed between three agents and I wanted to end the call well, so I told her the amount would be back with her in three days without checking the policy on partial refunds.

It was declined by my team leader the next morning. I asked to be the one to call her back, explained the error was mine rather than the company’s, and got approval for a smaller credit plus a waived fee, which she accepted.

The change I made is simple: I do not state an amount on a call unless I have the rule in front of me, and if I cannot find it, I commit to a callback time instead. It has cost me a few awkward silences and no retractions.

Student or first-job failure

My clearest failure was the group project in my final year, where I was the one who let it slip.

I had taken on the data collection for a four-person study and also had a part-time job going through its busiest month. I kept telling the group it was in hand because I expected to catch up at the weekend.

Two weeks before the deadline I had about half the responses we needed. I told the group, we cut the second research question, and I ran the remaining interviews in four days with help from one of them. We passed, with a weaker piece of work than we should have had.

What I took from it is that the cost of saying “I’m behind” early is almost nothing, and the cost of saying it late is the whole project. I now give people a status I would be comfortable having checked.

The follow-up questions

The first answer is rarely the end. Interviewers probe this one more than most, because it is where rehearsed answers come apart.

“What would you do differently?” wants a decision, not a sentiment. Name the point in the story where you would act differently and what you would do instead.

“Whose fault was it?” is a trap only if you are looking for somewhere to put the blame. Say what share was yours, acknowledge the rest without elaborating, and move back to your own part.

“Did you tell anyone at the time?” is the real question hiding inside the whole topic. If you told someone promptly, say when and whom. If you did not, say so and say what you do now; late disclosure honestly described is better than an invented early one.

“How did you feel about it?” invites one honest sentence. Uncomfortable, embarrassed, annoyed with yourself: all fine. Then return to what you did.

“What is your biggest failure?” asked after your example is usually a check on whether the first story was your real one. You can say that it is the one you have just described and why it stands out, or offer a second, shorter example. Having two prepared is worth the half hour.

Related wordings

“Tell me about a time you made a mistake” is narrower. A mistake is a specific error, often small, so the answer can be shorter and the correction beat carries even more of the weight.

“Describe a project that did not go well” allows a failure that was not entirely yours. You still need to identify your own contribution, but there is room to describe a plan that was wrong from the start and how you handled the consequences.

“Tell me about a time you received difficult feedback” is adjacent, and the structure holds: what was said, what you did with it, what is different now. Do not use a feedback story as a substitute for a failure story; interviewers notice the swap.

“What is your greatest weakness?” is a separate question with a separate answer, covered on the weaknesses page. The failure question wants one event; the weakness question wants a pattern you manage.

Finally, there is the deliberately broad version, “Define success. Define failure.” Answer the second half with a standard rather than an anecdote. Something like: failure is letting a problem stay invisible until someone else has to solve it. That sort of definition reads as a person who has thought about their own record, which is the whole point of the question.