Most interviews are conversations. They start with small talk, wander into whatever the interviewer finds interesting that day, and end with a gut feeling. Two candidates for the same role get asked entirely different questions, and then somebody tries to compare the results.
That doesn’t work, and it fails in a specific direction: it selects for people who are good at interviews rather than people who are good at the job. Those groups overlap, but not nearly as much as hiring managers assume.
An unstructured interview tells you how much you enjoyed talking to someone. That is real information — it is just not the information you were hiring for.
What “structured” actually means
A structured interview has three properties, and dropping any one of them collapses the whole thing back into a conversation:
- The same questions, in the same order, for every candidate. Not a rough theme list — the actual wording, fixed in advance.
- A defined scoring rubric per question. Written before the first interview, describing what a weak, adequate and strong answer contains.
- Scores recorded during the interview, not after. Memory reorganises itself around your conclusion within minutes of the call ending.
The rigidity is the point. Once every candidate has answered the same question against the same rubric, the comparison between them means something — and so does the comparison between what a candidate claimed and what they demonstrated.
The three inconsistencies it surfaces
1. Scope inflation
A résumé says “managed customer support for a SaaS company.” The structured question asks for the ticket volume, the tools, the escalation path, and who covered weekends. Candidates who owned the function answer in specifics without pausing. Candidates who assisted someone who owned it move to generalities. Neither answer is disqualifying on its own — but only one of them matches what the résumé implied.
2. Borrowed outcomes
Asking “tell me about a win” invites a team story. Asking “what specifically did you do, and what would have happened if you hadn’t been there” separates contribution from proximity. Strong candidates find this question easy. It’s a fair question and it isn’t a trap — it just requires that the story be theirs.
3. Communication that only works in short bursts
Written English and conversational English diverge sharply under time pressure. A structured interview holds every candidate at the same conversational depth for the same duration, which makes fluency measurable rather than impressionistic. This is why we run it after the English sample and before the practical test — the sample establishes a baseline, and the interview confirms it holds up live.
Where this sits in our process. The structured interview is check three of seven. It follows the application and the English sample, and it feeds the practical test — because a question the interview raises is exactly the thing the practical task should be designed to answer. Read the full seven checks.
What it does not catch
Being honest about the limits matters more than defending the method. A structured interview measures how someone describes and reasons about work. It does not measure whether they can do the work. Someone can explain inbox triage beautifully and still be disorganised in practice.
That gap is exactly why the interview is one check and not the whole process. The practical test measures execution. The role-play measures composure. The reference check measures whether the pattern held with someone who wasn’t in the room. Each check covers a blind spot in the one before it.
If you’re building your own process, start with the rubric rather than the questions. Write down what a strong answer looks like first — the questions worth asking tend to become obvious once you know what you’re listening for.