How to build one
- Start from the role, not from a library. List what the person must be able to do in the first six months and turn each into a criterion.
- Split criteria across stages. The intro call tests motivation and communication; the technical interview tests skill; the onsite tests collaboration and judgment. Nobody tests everything.
- Use four points. Strong no, no, yes, strong yes. An even scale removes the neutral middle that tells you nothing.
- Ask for evidence. One or two sentences per rating about what the candidate said or did.
- End with a recommendation and the single biggest concern.
Template: technical interview for an engineer
| Criterion | What a strong yes looks like | Rating |
|---|---|---|
| Problem solving | Breaks the problem down, states assumptions, checks edge cases without being asked | 1–4 |
| Code quality | Working, readable code; names things well; tests what matters | 1–4 |
| System thinking | Explains trade-offs between options and picks one for a reason | 1–4 |
| Ownership | Describes a project they drove from idea to production, including what went wrong | 1–4 |
| Communication | Thinks aloud, asks clarifying questions, takes a hint well | 1–4 |
| Overall | Would you be glad to work with this person next month? | Strong no / no / yes / strong yes |
Running the debrief
- Submit first, read later. Interviewers should not see other ratings until they have submitted their own, or the first opinion anchors the rest.
- Look at disagreements. Where one person said strong yes and another no, compare the evidence.
- Decide in the meeting. Hire, reject, or one named open question with an owner and a date.
A minute-by-minute agenda and a written version for async teams are in how to run an interview debrief.
Scorecards in Evander
Evander generates stages and a scorecard for each when you create a role, and you can edit them. An interviewer's view of other people's results stays hidden until they submit their own, and the debrief shows every interviewer's ratings side by side per criterion. A plain-English automation can remind interviewers who have not submitted after 24 hours. Scorecards are optional: a ten-person company can skip them for a role and still keep notes.
Questions
What is an interview scorecard?
A short form an interviewer completes right after an interview, with three to five criteria for that stage, a rating and evidence for each, and an overall recommendation.
What rating scale should an interview scorecard use?
A four-point scale: strong no, no, yes, strong yes. An even number removes the neutral middle option, which gives the hiring team nothing to act on.
How many criteria should a scorecard have?
Three to five per interview. Spread the criteria for the role across the stages so that each interview tests a few things well.
Should interviewers see each other's feedback before submitting?
No. Seeing another rating first anchors the interviewer. Collect each scorecard independently, then compare them in the debrief.