A single time limit applied to every question is the most common mistake in an async round, and it fails in both directions at once. Your warm-up question gets three minutes it does not need, so candidates pad. Your hardest question gets three minutes it cannot fit, so candidates get cut off mid-reasoning.
Rough starting points
- Warm-up or motivation: 60 to 90 seconds. Long enough to settle, short enough that nobody recites a cover letter.
- Experience and examples: 2 to 3 minutes. Enough for a situation, an action and an outcome.
- Technical or scenario walkthrough: 4 to 5 minutes. This is the one that needs room.
- Anything you want to ask us: 60 seconds, and treat it as optional.
These are starting points, not rules. The honest way to set a limit is to answer the question yourself, out loud, and then add about half again — you know the context and the candidate does not.
Why the limit helps the candidate
It sounds like a constraint imposed on them, and it is, but it also removes a decision. Without a limit, every candidate has to guess how much you want. The confident ones talk for eight minutes and the careful ones stop at forty seconds, and neither length tells you anything about their ability.
A stated limit is a shared expectation. It is one of the few places where a constraint makes an assessment fairer rather than harsher.
Watch for the cut-off cliff
If several candidates are being cut off on the same question, the limit is wrong, not the candidates. It is worth re-recording that round rather than penalising a group of people for your setting.
And on your side
Four questions at an average of two and a half minutes is a ten-minute review per candidate — and you can stop watching the moment you know. That is the number worth planning around, not the total recording length.
