The reason interviews feel unreliable is usually that nobody agreed what they were assessing. A scorecard fixes that, and it takes about an hour to write.
Start with outcomes, not attributes
A brief that asks for someone detail-oriented, proactive and a strong communicator has described almost every professional alive. It gives an interviewer nothing to test and nothing to disagree about.
Replace them with results. Month-end closes by the fifth working day. First response time under two hours. Pipeline reporting that the sales team trusts without checking. Now there is something to interview for.
Three outcomes, five competencies, one deal-breaker
- Three outcomes owned by day ninety, each measurable by someone other than the hire.
- Five competencies that produce those outcomes, ranked, so a trade-off is possible.
- One deal-breaker that ends the process regardless of everything else.
- A note on what this role is not, which prevents scope creep in the interview.
Ranking the competencies is the step people skip and the one that saves the argument later. When two candidates split the panel, the ranking decides it.
If two interviewers disagree, the scorecard was too vague, not the candidate too complicated.
Test the work, not the interview
A two-hour paid exercise on something real predicts performance better than a fourth conversation. Keep it small, pay for it, and grade it against the scorecard rather than against taste.
Have the technical stage run by someone who does the job. Founders assessing skills they do not hold tend to reward confidence, which is exactly the trait that survives a bad hire longest.
Use it after the hire too
The scorecard becomes the first thirty and ninety day plan, which means the new person starts against the same standard you selected them on. Most onboarding problems are just a scorecard that was thrown away at offer stage.
