Hiring Debrief From Interview Scorecards

Ask each interviewer to write up their notes and ratings alone, soon after the interview, and submit them before the debrief. In its guide to structured interviews, the US Office of Personnel Management (OPM) says each panelist should form an independent evaluation without discussion with other panel members¹. Only then does the panel compare notes.

Keep scorecards hidden from other interviewers until all of them are in. Greenhouse, which sells hiring software, studied scorecards from more than 10 million panel interviews. Interviewers who peeked at a colleague's scorecard before submitting were 3.6% more likely to give the exact same rating². Each rating that drifts toward a colleague's hides a disagreement the debrief should hear.

Set a deadline, such as the end of the next working day, and chase anyone who misses it. Notes written after the meeting record what the room decided, not what the interviewer saw.

Before the meeting, mark each comment as an observation or a verdict. An observation records the question and what the candidate said or did: "Asked who set the go-live date, she said the client's IT lead did." A verdict is a conclusion with nothing behind it, such as "not strategic," "lacks presence" or "great culture fit."

Only observations can be checked against someone else's notes. OPM's guide wants every rating "specific, defensible, and supported by behavioral examples," including actual examples of the candidate's answers¹. When a verdict matters, ask the interviewer what they saw. If nothing comes back, it stays on the record but carries no weight.

Some comments must not count at all. Strike any remark about age, accent, family plans, religion, health or appearance, and build no argument on it. In the US, the EEOC says an employer may not base hiring decisions on stereotypes and assumptions³ about protected traits. These include race, color, religion, sex, national origin, age (40 or older), disability and genetic information.

List the four to seven competencies the role needs, taken from the job description rather than invented in the room. Google's re:Work hiring guide recommends defining specific attributes for the role instead of relying on "gut feel" or vague cultural fits⁴.

Comparing one competency at a time keeps the discipline of a structured interview, and that discipline pays off. In a 2022 reanalysis of earlier hiring studies, Paul Sackett and colleagues found structured interviews were the strongest predictor of job performance⁵ among the methods they compared. Sorting also blunts the halo effect, where a strong impression on one skill lifts the ratings on the others.

File every comment under the competency it gives evidence for, with the interviewer's name beside it. Comments that fit nowhere go in their own pile, counted rather than deleted. The finished grid shows coverage: how many interviewers produced real evidence on each competency. It's often lopsided, with several people probing the same strength while a skill the job depends on goes untested.

Give the meeting a facilitator who isn't the hiring manager, and book 30 to 45 minutes. At Amazon, a trained employee from outside the team, called a Bar Raiser, facilitates the debrief to keep it data-driven⁶. One hiring manager there says it helps eliminate "the habit of looking to the most senior person in the room to make the decision."

1. 1

   Read the recommendations first

   The facilitator reads each submitted recommendation aloud before any discussion, so nobody shifts theirs to match the room.

2. 2

   Walk the grid one skill at a time

   For each competency, only the people who tested it share their evidence. Untested skills go on the gap list.

3. 3

   Test each disagreement

   Ask whether both people tested the same skill. If not, record two findings. If so, compare their observations.

4. 4

   Find the question it turns on

   For each real conflict, write the one question that would settle it, who can answer it and by when.

5. 5

   Decide, or list what's missing

   Decide only if the observations support it. Otherwise, name what's outstanding and when it will be closed.

Don't paper over a split with an average. On a 4-point scale, a 4 and a 1 average 2.5: a score neither interviewer gave, which buries the disagreement worth resolving. Talk it through instead. OPM's guide asks panelists to thoroughly explore the basis for discrepancies in their ratings¹, then reach a consensus.

A real conflict often turns on a single fact. If two interviewers heard different accounts of who made a go-live call, one reference question can settle it. An untested competency needs a short test of its own before anyone decides. A 20-minute exercise often does it, or a call built on one question from a behavioral interview question bank.

Then write down the decision and what carried it: the deciding observations, how each conflict was settled, and anything still untested. In the US, covered employers must keep records having to do with hiring for a year⁷ after they are made or after the decision, whichever comes later. If someone files a discrimination charge, keep the relevant records until it is resolved.

Short on time before the debrief? The tool at the top of this page sorts your panel's comments into this grid and drafts the write-up. For the next candidate, the interview process and scorecard template gives every competency one interviewer to own.