“Tell Me About a Time You Solved a Problem”
A hard task you finished isn't a problem you solved — whoever assigned it already knew what was wrong. What they're really asking is whether you can find a cause nobody has handed you, and if you're changing careers, whether you think or just execute. Get an answer built on a problem nobody could explain, the wrong assumption you started with, the unglamorous way you found the real reason, and how you checked weeks later that the fix held.
Generated for: Auto parts store assistant manager (5 years) interviewing for a supply chain coordinator role
For about four months our inventory counts came up short on the same handful of fast-moving parts — filters, wiper blades, that tier. Not huge numbers, but every single month, and always the same SKUs. The store consensus was theft, and honestly that was my assumption too. We'd already moved two of those bins behind the counter and it hadn't changed anything, which should have told us something earlier than it did.
So instead of tightening security again I went and looked at the boring end of it. I pulled the receiving paperwork for those parts going back three months and sat in on two deliveries myself, which I'd never actually done — I'd always signed for them from the office. On the second one I caught it: one of our vendors had switched from a 12-count case to a 10-count and kept the same part number, so every time we received a case we were booking two units we never had. The guy on our dock had noticed the cases felt light and mentioned it to somebody months earlier, and it just never got to anyone who could act on it.
The fix was small — receiving physically counts that vendor's cases instead of scanning the case label, and I flagged the SKU with our rep. I waited for the next two monthly counts before I believed it, and the variance on those parts went to basically zero. I'll be honest that it didn't fix everything; we still had shrink on the small electrical stuff, and that one probably was theft. What it changed for me is that when everybody already agrees on the cause, that's usually the thing worth checking first.
Why this one works
- The cause is genuinely unknown at the start, and the candidate says out loud that they believed the wrong explanation everyone else believed — which is what makes the diagnosis credible instead of lucky.
- The method is concrete and unglamorous: pull the paperwork, stand on the dock, listen to the person doing the job. A listener can picture it, and it transfers directly to an office analysis role.
- The fix is verified across two later counts rather than declared solved, and the answer admits a piece that stayed broken instead of overclaiming a company-wide win.
Yours is written from your own background and the role you're actually interviewing for — not this one. One credit per answer, $9.99 for a single credit or $49.99 for ten. Fill in the form below to generate it.