Guides
How to build an interview scorecard that produces fair, comparable decisions
Without a scorecard, interviews collapse into gut feeling and whoever spoke last. A simple, criteria-based scorecard makes hiring fairer, more defensible, and far easier to compare.
Key takeaways
Take criteria from the intake
4 to 6, mapped to must-haves + success picture.
Write a 1 to 4 scale with anchors
Describe what each level looks like.
Score independently
Each interviewer rates before any group talk.
Compare + decide on evidence
Discuss the disagreements; decide from the scores.
1. Why a scorecard beats a gut verdict
A scorecard forces interviewers to rate a candidate against the specific things the role needs, instead of collapsing the whole interview into a single "yeah, I liked them". That single impression is exactly where bias and recency live, the confident talker, the person who reminds you of yourself, the strong finish that erases a weak middle. Scoring each criterion separately makes the decision legible, comparable across candidates, and defensible if it's ever questioned. In Germany it's also a core part of an AGG-compliant, bias-resistant process.
Score evidence, not personality
Anchor every criterion to something the candidate did or demonstrated, not to traits like "culture fit" that quietly become "is like us". Evidence-based scoring is both fairer and more predictive, and it keeps the process defensible under the AGG.
2. Derive the criteria from the intake
Don't invent generic criteria. Take them straight from the must-haves and success picture you agreed in the intake meeting, that's the whole point of having one. A scorecard with four to six criteria that actually map to the role (a specific skill, a relevant kind of experience, a way of working the team needs) beats a long list of vague traits like "communication" that everyone scores differently. Each criterion should be something you can gather real evidence for in the conversation, not a personality guess.
3. Use a scale with behavioural anchors
A 1-to-4 scale works well (an even number quietly forces a lean rather than a safe middle). But numbers alone drift, your 3 isn't my 3. Add a short behavioural anchor to each level: what does a 2 look like versus a 4 on this criterion? Anchors turn a subjective number into a shared standard and make scores comparable across different interviewers. Capture a sentence of evidence next to each score ("rated 4: walked through a near-identical migration they led"), so the rating is grounded, not just asserted.
4. Score independently, then compare
The order matters. Each interviewer fills in their scorecard BEFORE the group discusses, because the moment a senior voice says "I loved them" out loud, everyone's scores quietly converge on it, that's groupthink erasing the value of multiple perspectives. Independent-first, discuss-second surfaces real disagreement, which is often the most useful signal you have. In KI BMS the scorecard lives on the application and feeds the candidate's fit picture, so the structured evaluation you designed is what actually drives the decision and stays on the record, rather than a hallway consensus nobody can reconstruct later.
FAQ
Frequently asked
Try KI BMS
Free plan, no credit card. We host in Germany. You can export and delete everything self-serve.
Read next
The phone screen: a 15-minute guide that saves you hours of interviews
A short structured first call that filters before the expensive interviews.
Read
How to build and maintain a talent pool instead of starting every role from zero
Turn near-miss candidates into a warm shortlist for your next opening.
Read
How to set up an employee referral programme that actually brings applications
Make referring effortless and rewarding it visibly fair.
Read