Culture Interview

There's a pattern where the skills check out and the person leaves after three months. The code was good, the experience matched, and then the friction started. A decision gets made in a meeting, the new hire goes back unconvinced and builds it another way, and when that comes up they ask why nobody said so earlier. Neither side is saying anything false, and the same wall shows up every week.

The cause is usually not ability but a difference in organizational culture — the shared habits and attitudes about how work gets done. Someone from a company that writes decisions down before acting and someone from a company that settles things verbally and starts immediately will remember the same meeting differently. Teams that give feedback on the spot and teams that pull you aside afterward produce the same mismatch. Neither is correct; they're just different, and a technical interview never surfaces any of it.

A culture interview exists to surface it before the offer. The goal isn't to find someone who fits perfectly, but to let both sides see where they won't fit before anyone commits. Where the rest of the process asks what a candidate can do, this stage asks what they habitually do.

Decide what you're checking, turn it into questions about past behavior, have interviewers score separately, and only then compare. Five steps.

1. Write the standard as a sentence. A value like "candor" can't be used as written, so it becomes something observable: "says they disagree before the decision, not after." Three or four of these is enough.
2. Put the questions in past tense. "How would you handle a conflict?" gets a model answer from anyone. "Tell me about a time you disagreed with a team decision" pulls out something that actually happened.
3. Give it 45 minutes to an hour, and leave 15 of those for the candidate's questions. The session is also where the company explains itself, so if only the interviewers talk, half of it was wasted.
4. Have each interviewer record a score and a reason on their own. Discuss first and the loudest opinion in the room pulls the rest with it. Compare after submission and you can see exactly where the readings diverged.
5. Write down the decision and the reason for it. An interviewer voting no records which answer, and which part of it, drove that. A note that only preserves a feeling is useless for the next candidate, and it's also the only thing that makes two candidates from different months comparable.

Walk in without a written standard and the session turns into small talk. The requirements on a job description are what make a technical interview consistent; without an equivalent list of behaviors here, each interviewer measures something different and objects for a different reason.

Two or three people from outside the hiring team, at the point where the role-specific checks are finished and the offer hasn't gone out yet.

Outsiders are there for two reasons. People on the hiring team lean toward filling the seat that's urgent right now, and the way the company works isn't the property of one team anyway. The hiring manager — the person who needs the role filled and makes the final call — usually steps back and receives the result instead. Someone running the last check on a candidate they already want isn't running a check. Teams that skip this separation tend to notice it only in hindsight, when every recent hire came through the same person's judgment.

The order is resume screening, then the technical interview and any take-home assignment, then this stage, then a reference check, then the offer. Move it earlier and three people spend an hour on candidates who were going to fail the role-specific bar anyway. Move it after the offer and reversing costs far more: it's awkward to schedule another interview with someone you've already discussed a start date with, and objections raised at that point tend to get buried.

"She's through on the role side. We just need to schedule the culture interview."

The skills are settled; what's left is how the person works.

"There's a no vote, but the reason is 'something felt off.' We need to go back for an actual basis."

If the record has no behavior in it, it can't be used as a decision.

"Can someone from the other team sit in this round?"

Only having the hiring team in the room means nothing is really being checked.

At Pagedeck, Grace used to take these herself, and past forty people the company switched to rotating one interviewer from each team. Having the founder in every interview stretched the schedule, and it also made the outcome look, from the candidate's side, like one person's opinion.

In a round Mia sat in on, one objection came back. Not that the candidate was unpleasant, but that when asked what they did the last time a deadline slipped, the answer was that they finished it alone over the weekend. Pagedeck's standard is to raise a slip early, so that gap got discussed before the start date. The objection didn't sink the hire; it meant the first week started with one thing already settled.

The most common reading is that this measures personality. The word culture is in the name and the conversation is relaxed, so it feels that way. What's actually under examination is behavior — what someone does when opinions split — not taste. Once taste drives the decision, the people who are easy to have lunch with get hired, and within six months everyone left has a similar background and a similar way of thinking. That team is comfortable because nobody objects in meetings, and it misses the same things together. The fix isn't to lower the bar but to name the specific behaviors being checked, so that "different from us" and "won't work the way we work" stop being the same verdict.

The other mistake is treating it as a formality at the end. The skills were verified upstream, so this looks like a ceremony, and with no preparation it becomes one: 30 minutes of chat producing an impression, and an impression with no basis is impossible to argue either way. When the stage is hollow, the problems it was meant to catch pass through and appear after the start date, coming back as early attrition. One person leaving at month three means paying for the whole hiring loop again.