An interview scorecard that survives a debrief

Most scorecards fail for the same reason: a five point scale with no definitions, so a 3 means whatever the interviewer felt that afternoon. Fix the scale and the definitions and the debrief gets twenty minutes shorter.

Start for free Book a demo

$100 of credits on us. No card required.

Score each attribute separately, write the evidence before the number, and never average the scores into a single figure. The averaging is where good candidates die.

Interviewers submit independently before the debrief. If people see each other's scores first, you are running a consensus exercise rather than an evaluation.

Copy the one you need

1. The scale, defined

Paste this at the top of every scorecard

4 · Strong yes. Clear evidence they are above the bar. I would fight to hire them.
3 · Yes. Evidence they meet the bar for this attribute. No serious concerns.
2 · No. Some evidence, not enough for this level. Would be a stretch.
1 · Strong no. Evidence they are below the bar, or a value conflict.

No 5 point scale, because the middle option is where undecided interviewers hide.
Every score needs one line of evidence. A score without evidence does not count in the debrief.

2. Attribute scorecard

One per interviewer, per interview

Candidate: [Name]
Role: [Title]
Interviewer: [Name]
Interview: [Stage, e.g. technical deep dive]
Date: [Date]

Attribute 1: [e.g. Systems design depth]
Score: [1-4]
Evidence: [What they said or did. Quote where you can.]

Attribute 2: [e.g. Ownership]
Score: [1-4]
Evidence:

Attribute 3: [e.g. Communication under ambiguity]
Score: [1-4]
Evidence:

Attribute 4: [Role specific]
Score: [1-4]
Evidence:

Overall recommendation: [Strong yes / Yes / No / Strong no]
The one thing that would change my mind: [Write this even when you are certain.]
What I did not get to test: [The gap the next interviewer should cover.]

3. Debrief format

Run this in 25 minutes, not an hour

1. Everyone states their overall recommendation in one word. No reasoning yet. (2 min)
2. Lowest scorer speaks first, with evidence. (5 min)
3. Highest scorer responds, with evidence. (5 min)
4. Open discussion, evidence only. Opinions without evidence get named as such. (8 min)
5. Hiring manager decides. Not a vote, a decision, on the record. (5 min)

Write the decision and the reason in the ATS the same day. Six weeks from now the reason is the only part anyone needs.

4. Take home review

Async work sample

Candidate: [Name]
Time they reported spending: [hours]
Reviewer: [Name]

What the brief asked for: [one line]

Did it work: [yes / partly / no, plus what broke]
Code or craft quality: [1-4] · Evidence:
Decisions and trade offs: [1-4] · Evidence: [What did they choose not to do, and did they say why?]
Communication in the readme: [1-4] · Evidence:

Would I want this in our codebase: [yes / no]
Question to ask in the follow up: [The interesting choice they made.]

Where scorecards usually go wrong

Averaging attribute scores into one number hides the pattern. A candidate at 4, 4, 4, 1 is a different conversation from four 3s, and the average is the same.

Scoring after the debrief instead of before it turns the process into whoever spoke most confidently.

More than five attributes and interviewers stop reading them. Pick the ones that predict success in this role, not a company wide list.

The same scorecard for every level. Senior and junior need different definitions of the same attribute or the scale means nothing.

Two interviewers scoring the same person should reach the same number.

When they do not, the disagreement should be about evidence rather than about what a 3 means.

I score the same way: one bar, agreed once, applied identically every time.

What people ask about this

Four point or five point scale?

Four. Removing the neutral middle forces a position, which is the entire point of asking. Undecided interviewers should write what would change their mind instead of parking on a 3.

Should interviewers see each other's scores before the debrief?

No. Independent submission first. Otherwise the first score anchors the room and you get agreement instead of information.

How does Laidback use scorecards?

I read the attributes on your existing scorecards to build the brief, and I calibrate my scoring against the candidates your team actually advanced. The bar comes from your decisions, not from a generic model of a good hire.