Interview scorecard

Scorecards exist because unstructured feedback is unreliable. Two interviewers who meet the same candidate will often describe the conversation differently, and neither description can be compared with the next candidate's. A scorecard fixes the criteria before the interview, so each interviewer assesses the same things, and asks for evidence, so a rating of four out of five for stakeholder management comes with the example that justified it.

For an agency, the scorecard is also how client feedback reaches the record. Hiring managers are slow to give feedback by email and vague when they do. A short form with the role's criteria already on it, sent straight after the interview, gets a faster and more specific answer, and that answer is what the consultant needs to prepare the candidate for the next round or explain a rejection.

A common shape is a one-to-five rating per criterion and a five-point overall recommendation: strong yes, yes, maybe, no, strong no. The recommendation scale matters more than it looks. If one interviewer's scale runs from hire to no hire and another's from one to ten, the two cannot be combined into a decision.

Scorecards are only as good as the criteria on them. The best are drawn from the brief: the three or four things the hiring manager said were essential, each with a short description of what good looks like. Generic criteria such as communication and culture fit invite generic ratings. Where a client runs several stages, each stage should assess something different, so the second interview is not a repeat of the first with a different person asking.

Where it goes wrong: criteria that differ from the brief; scorecards filled in days later from memory; one interviewer's view recorded and the other two lost; and rejection reasons that never make it onto the candidate's record, so the next consultant has no idea why they were turned down.