The structured interview scorecard: building one for a specific role

Echipa HR 365 · reviewed 2026-09-02 · 5 min read

A structured interview scorecard means: the same competencies assessed for every candidate, the same questions asked in the same order, and a scale with written descriptions for each level. The effort is in building it; after that, every interview becomes comparable with the others, which a free-form interview never allows.

Why structure beats impression

In a free-form interview, each interviewer explores something different, and at the end the comparison is between impressions formed in different conversations. Research on selection methods has shown for decades that structured interviews predict job performance visibly better than unstructured ones, and yet the second remains the de facto standard, because it is more comfortable and feels more natural.

The second argument concerns the variability of ratings: two people hearing exactly the same answer rate it very differently if they have no written description in advance of what each level means. The scorecard does not make them think alike — it makes them talk about the same thing.

Step 1: choose six competencies, not twelve

Six is the practical limit for a one-hour interview: each competency needs a main question, one or two probes and time for a concrete answer. Beyond six, you either cut the probing and rate impressions, or you overrun the hour.

How you choose them

1. Manager — describes three real situations from recent months where somebody in this role made the difference, for better or worse
2. Recruitment — extracts what was decisive — not generic labels, but what the person actually did
3. Manager and recruitment — cut the list to six, removing anything easily learned in the first months
4. Manager — marks which two are knock-outs, so the final decision is not taken on an average

Step 2: write the questions about facts, not hypotheticals

A good question asks for a lived situation, not an opinion. “How would you handle a client complaint” invites theory, and anybody who has read something answers well. “Tell me about the last time a client complained and what you did” asks for a real case, with details you can probe.

For each main question, prepare two probes: “what did you do, specifically” and “what happened afterwards”. The first separates their real contribution from the team’s story, the second checks whether the result actually existed.

Step 3: the scale, with descriptions, not numbers

A 1-to-5 scale with no descriptions produces ratings that mean something different to each assessor. Four described levels are enough and remove the comfortable middle where half the ratings otherwise land.

The completion rule: every rating has a sentence of evidence beside it — what the candidate said, not what the assessor felt. A scorecard filled in with numbers alone cannot be discussed at the end of the process.

Step 4: decide before comparing

Every assessor completes their scorecard independently, before the joint discussion. If ratings are shared during the discussion, the first opinion voiced moves the others, and four assessments become, in practice, one. It is the cheapest possible correction and almost nobody makes it.

How long a scorecard takes and how many you need

The first scorecard for a role family takes about two hours of work with the manager — one hour for the real situations and the competencies, one for phrasing the questions. Subsequent ones take thirty minutes, because the scale and the structure are already fixed.

You do not need one scorecard per post, but one per role family: every operational post in the same department can share the same six competencies, with one or two swapped. A mid-sized company works well with five to seven scorecards in total, reviewed once a year.

Does it not make the interview rigid and unnatural?

The structure is in the main questions and the scale, not in the conversation. The probing stays free; what gets fixed is what you assess and how you rate — exactly the part where improvisation helps nobody.

Do I use the same scorecard for every role?

The scale yes, the competencies no. A shared scale makes ratings comparable across roles; shared competencies would mean assessing the same thing in an accountant and a salesperson.