D · Faster, fairer hiring
Interview scorecards your hiring managers will actually use
A copyable interview scorecard: five criteria, a four-point anchored scale and an evidence box. Structured interviews carry the highest measured validity at .42.
- An interview scorecard that gets used has five criteria or fewer, one question per criterion, a four-point scale with written anchors, and a box for the evidence behind each score. Interviewers submit it before the debrief starts. The template below is ready to copy, and the rules under it are the part that keeps it alive after the second week.
Why the structure earns its overhead
- SHRM defines a structured interview as one in which candidates are asked the same predetermined questions in the same order, with all responses evaluated using the same rating scale. Its benchmarking data shows how uncommon that still is: structured interviews were used to assess executive candidates in 34% of organisations, middle management in 37%, and individual contributors in 36%, against in-person interviews at 79%, 78% and 76%.
- The payoff is measurable. Sackett and colleagues' 2022 re-analysis of the selection literature puts structured interviews at the top of the validity table with a mean of .42, ahead of job knowledge tests at .40, empirically keyed biodata at .38, work sample tests at .33 and cognitive ability at .31. The same analysis reports an 80% credibility interval of .18 to .66 for structured interviews, so the design of the questions and the scoring decides where inside that spread an employer lands. A badly built scorecard gets you the bottom of the range and the overhead of the top.
- The UK picture puts the gap in execution. In the CIPD's 2024 survey of 1,016 organisations, 73% reported that line managers follow objective assessment and scoring criteria when recruiting, up from 67% in 2022, while only 28% train all interviewers on their legal obligations and objectives.
The scorecard
- Swap the first criterion for the technical core of your role and keep the rest. Each interviewer owns two or three rows for one candidate, asks the questions for those rows, and scores nothing else.
| Criterion | What a strong answer contains | Question that produces it |
|---|---|---|
| Core skill for the role | A piece of work they owned from start to finish, the constraint they worked under, and the measurable result | Walk me through the last time you did this work. What was the constraint, and how did it end? |
| Judgement with incomplete information | A decision made without the data they wanted, the trade-off they accepted, and what they would change now | Tell me about a decision you made without enough information. What did you trade away? |
| Working through disagreement | A named disagreement, the other person's position stated fairly, and how it resolved | Describe a time a colleague disagreed with your approach. What was their case? |
| Ownership of a failure | A failure they caused, what they did in the following week, and the change that stuck | Tell me about something you got wrong that cost your team time. |
| Motivation for this role | Reasons drawn from the work described in the job brief. Company profile alone scores low | What in this job description made you apply? |
The scale
| Score | Anchor |
|---|---|
| 4 | Evidence beyond what the role requires. You would hire this person for a harder version of the job. |
| 3 | Evidence meets the requirement. You can name the example that proves it. |
| 2 | Partial evidence. Something adjacent, or one thin example. |
| 1 | No evidence. The answer stayed general or moved to another topic. |
- Four points, no midpoint. On a five-point scale interviewers park uncertain candidates on the 3 and the whole loop returns the same score for everyone.
The evidence box
- Under every score, one line naming what the candidate actually said. "Rebuilt the billing reconciliation job after a duplicate-charge incident, cut the manual review from 4 hours a week to 20 minutes" is evidence. "Strong technically" is a feeling with a number attached to it.
- A score arriving without that line goes back to the interviewer before the debrief opens.
Six rules that keep it in use
- Collect scores before the debrief, in writing, with no discussion first. A debrief that opens with a spoken opinion has already set the room, and the written scores that follow will agree with it.
- Give each interviewer two or three criteria, never all five. Overlap two of them deliberately across two interviewers so you can see whether they agree.
- Ask the same questions in the same order for every candidate on the requisition. Changing the question changes what the score means, and the comparison is the product.
- Delete the row called culture fit. It has no anchor anyone can write down, and it scores familiarity. If there is a behaviour you genuinely need, name the behaviour and write an anchor for it.
- Rewrite the criteria per requisition, from the job brief, at the same meeting where you agree the headcount. A scorecard inherited from a different role scores the previous job.
- Keep the scores. They are the input to [measuring quality of hire](/spaces/how-to-measure-quality-of-hire): 51% of the organisations that measure it compare interview scores against later performance, and that comparison needs both halves stored against the same person.
What it does to your timeline
- Structured scoring shortens the loop. Each round returns a number that can be compared against another candidate's, so the fourth interview stops being the one that exists because nobody could articulate their reservation about the third. SHRM's benchmarking data gives the interview stage a 7-day median and a 9-day average, and every extra round added after an inconclusive debrief lands inside that stage. [How to reduce time-to-hire](/spaces/how-to-reduce-time-to-hire) covers the rest of the stage data.
- The scorecard also only ever sees the people who reached the interview. If your screen filtered on CV keywords, the loop is scoring a pool that already lost qualified applicants, which is the failure described in [why an ATS rejects qualified candidates](/spaces/why-ats-rejects-qualified-candidates) and the reason for [hiring on proof of fit](/spaces/hiring-on-proof-of-fit) at the screening stage.
- In SHRM's data, 13% to 14% of organisations still run explicitly unstructured interviews. The scorecard above takes about an hour to write for one role.
Frequently asked questions
- What is an interview scorecard?
- A fixed list of criteria for one role, each with a question that produces evidence and an anchored rating scale that every interviewer applies the same way. SHRM's definition of a structured interview requires the same predetermined questions in the same order, with every response evaluated on the same scale.
- How many criteria should an interview scorecard have?
- Five or fewer per interviewer. Longer scorecards get filled in from memory after the fact, which removes the comparability that made them worth having. Split the criteria across the interview loop so each interviewer owns two or three and asks the questions for those.
- What rating scale works best for interview scoring?
- A four-point scale with written anchors and no midpoint. Every point should describe the evidence that earns it, so a 3 means the interviewer can name the example that proves the requirement. An odd-numbered scale gives interviewers a neutral option, and it is the one they take when the evidence is thin.
- Do structured interviews actually work better?
- The 2022 re-analysis of selection research by Sackett and colleagues puts structured interviews at the top of the validity table with a mean of .42, ahead of job knowledge tests at .40 and cognitive ability tests at .31. The same work reports an 80% credibility interval of .18 to .66, so the design and the discipline around it decide the result.
- Why do interview scorecards get ignored?
- Usually because scores are collected during the debrief rather than before it, so the first person to speak sets the room. Collect every score in writing before anyone shares an opinion, and reject a score submitted without a line of evidence behind it.
Sources
Next step
Try it on your next role
Unframed builds shortlists by ranking people on proof of fit.