Behavioural interviews · Module 06 of 12 · Level 2: The core questions

Failure, mistakes and feedback

"Tell me about a time you failed." Most candidates dread it and reach for a fake failure. Interviewers want a real one, because what they're scoring isn't the mistake. It's how fast you owned it, how you fixed it, and what you changed so it couldn't happen again.

1. The idea

Airline pilots are trained to report their own mistakes, and the industry studies them, because every reported near miss makes the next flight safer. Nobody thinks less of a pilot for the report; they'd think much less of one who hid it. Failure questions test whether you work like that.

Two related question families:

Failure and mistakes: "Tell me about a time you failed", "a mistake you made", "a project that didn't go to plan", "a time you missed a deadline".
Feedback: "The hardest feedback you've received", "a time you changed based on feedback", "feedback you gave someone".

Both want the same arc:

A real failure, and it was yours. Not a disguised strength, not someone else's fault. It cost something: time, money, users, trust.
You owned it quickly. You told the people affected before they found out, without excuses.
You fixed what you could. The immediate damage, first.
You changed something so it wouldn't happen again: a check, a process, a habit. For yourself, and ideally for the team.
You can show the change working. "Since then…" is the most valuable sentence in the answer.

For feedback the arc is the same: you heard it without defending yourself, made sure you understood it, changed something specific, and checked back later.

2. How it works

One mistake, told the way strong failure answers go: the damage, the ownership, the fix, and the change that outlived it.

3. Try it yourself

Would you tell this failure?

Six candidate answers to "Tell me about a time you failed." Decide which ones to use in an interview, then check.

4. What interviewers look for

Failure answers are scored on honesty and on what changed afterwards. The failure itself is the least important part.

Signal	Weak	Good	Strong
Realness	A disguised strength ("I care too much")	A real but small mistake	A real failure with a real cost, clearly yours
Ownership	Blames others or circumstances	Accepts it	Told the people affected first, without excuses
Fix	Not mentioned	Fixed the immediate problem	Fixed it, then found the root cause
Lasting change	"I learned to be more careful"	A personal habit	A change to how you or the team works, with evidence it holds

5. A model answer

"Tell me about a time you made a mistake." An invented answer, shown part by part. Read it for the shape, then build your own from a true story.

SituationLast year I ran a database migration that added an index to our orders table during working hours.

TaskIt was my change, and I'd tested it on a staging database a tenth the size of production.

ActionThe index build locked the table for nine minutes, and checkout failed for every customer during that time. I saw the alerts within a minute, cancelled the migration, and posted in the incident channel that it was my change and what I'd done. Once checkout recovered, I wrote the incident review myself. The root cause wasn't just my test: our migration tool didn't warn about locking operations at all. I added a check to CI that flags any migration that would lock a large table and requires it to run in the off-peak window, and I re-ran my index online that night.

ResultWe lost roughly 600 orders' worth of checkout attempts, and most customers retried. Since the CI check went in, it has flagged 11 risky migrations from five teams, and we haven't had a migration lock since.

LearningI now ask "what does this do at production size?" before any schema change, and I test on a production-sized copy when the table is big.

Why it scoresThe failure is real and has a cost. The candidate owns it within minutes, publicly. The fix goes past their own mistake to the system that allowed it, and there's evidence (11 flagged migrations) that the change outlived the incident.

6. Remember it

Memory hook: "Own it, fix it, change it"Admit it quickly, repair the damage, then change something so it can't happen the same way again.

How this hook works: what “Own it, fix it, change it” means

What the slogan means, phrase by phrase

“Own it”: Say it was yours, early, to the people it affected, without excuses.
“fix it”: Stop the bleeding first: the immediate damage.
“change it”: Find the root cause and change a check, a process or a habit, then show it working since.

Why this picture fits: Pilots report their own near misses so the whole industry can change a checklist. Owning it is the report; fixing it is landing safely; changing it is the new checklist line.

The story → the concept

The flight	Your failure
The near miss	The mistake
Filing the report	Telling people it was yours
Landing safely	Fixing the immediate damage
The new checklist line	The CI check, process or habit that prevents a repeat

Watch it on an example

Own it: "I posted in the incident channel that it was my migration."
Fix it: "I cancelled it; checkout recovered in nine minutes."
Change it: "I added a CI check for locking migrations; it has caught 11 since."

The trap it saves you from

Picking a fake failure to look good, which makes you look evasive instead.

It still works when…

The failure was a team's, not yours alone. Then own your part of it specifically, and describe the change you drove.

7. Common mistakes

The humblebrag"I work too hard" or "I'm a perfectionist" tells the interviewer you won't admit real mistakes.

Someone else's faultA failure story where the cause was another team is a complaint, not a failure story.

A catastrophe with no fixPick a real failure, but one you recovered from and learned from. The story is about the recovery.

A vague lesson"I learned to communicate better" is not a change. Name the specific thing you do differently now.

Defending yourself on feedbackIn feedback stories, don't relitigate whether the feedback was fair. Show what you did with it.

8. Practice exercise (25 minutes)

List three real failures (5 min). Ones that cost something, and were at least partly yours.
Pick the one with the best change (5 min). The one where you changed something that still holds today.
Write the arc (10 min). Own it, fix it, change it, and one "since then" sentence with evidence.
Add a feedback story (5 min). Hard feedback you got, what you changed, and how you checked it worked.

9. Self-check quiz

Which failure makes the best interview story?

Real, owned and followed by a lasting change.

What's the most valuable sentence in a failure answer?

Evidence that the change you made holds.

You received feedback you thought was unfair. How should the story go?

The interviewer is testing how you take feedback, not whether it was fair.

Which lesson is specific enough?

A concrete change someone could observe.

10. Go deeper

Resource	Time	Why
Postmortem culture: learning from failure (Google SRE book)	15 min	Blameless incident reviews: the professional version of "own it, fix it, change it".
Growth mindset (Microsoft culture)	5 min	"Taking smart risks and learning from mistakes is how we move forward."
Module 07: Leadership and influence	Next	Leading without the authority to order anyone.