A scorecard's job is not to record an opinion. It is to make a debrief tractable — to turn "I liked them" and "I didn't" into a disagreement about something specific enough to resolve. Most fail at this, and they fail in predictable ways.
Failure one: unanchored scales
A 1-to-10 with no descriptions is not a scale, it is a mood. One interviewer's 7 is another's 5, and the debrief becomes a negotiation about numbers rather than about the candidate.
Write one sentence per level, per criterion. Not all ten — three is enough. What a 4 looks like, what a 7 looks like, what a 9 looks like. Interviewers will still disagree; they will disagree about evidence.
Failure two: criteria that measure the same thing
"Communication", "collaboration" and "stakeholder management" on one scorecard will move together on almost every candidate. You have not measured three things, you have measured one thing three times and given it triple weight without deciding to.
Four to six criteria that genuinely pull apart beats ten that correlate.
Failure three: scores collected in the room
If people score after discussing, you have one opinion and three echoes. The most senior voice anchors everyone else within about ninety seconds, and the scorecard records the anchoring rather than the assessment.
Independent scores first, submitted and locked. Then discuss. The disagreements that survive that process are the informative ones.
What holds up
The scorecards that still make sense a month later share three properties:
- Anchored — every level has a description someone else would recognise
- Evidenced — each score carries the moment that produced it, quoted
- Independent — recorded before anyone spoke
The third is free. The first takes an afternoon per role. The second is the one worth building tooling for, because asking interviewers to transcribe their own justification is asking for the thing they will skip when they are busy.
The test
Take a scorecard from a hire you made six months ago. Can you tell, from the document alone, why that person was hired over the runner-up?
If not, the scorecard was a formality. That is worth knowing before the next one.