Candidate assessment on a small hiring desk is usually three people forming an impression in three separate conversations and then arguing about it in a fourth. That is not a bad process because the people are bad at judging; it is a bad process because nothing was written down before the interviews that says what was being judged or how much each part counted. The fix is small and unglamorous: agree the criteria and their weights before anyone meets a candidate, score each candidate against each criterion on the same scale, and let the arithmetic produce the ranking. This page sets out what to put on the sheet, what the weights actually do to the answer, and where the free scoring sheet on this site works it from your own numbers.
Agree the criteria and the weights first
Weights agreed after the interviews describe the interviews. Agree them with whoever owns the role, before anyone meets a candidate, and write down what a 3 and what an 8 look like for each criterion. This is the whole trick: an unanchored ten point scale is a mood, and two interviewers using it will disagree about a candidate they in fact agree about. Three to five criteria is usually enough, because a longer list dilutes the ones that decide the job into the noise of the ones that do not.
Score each candidate on the same scale, separately
Have each interviewer score before the debrief rather than during it, because the first opinion spoken in a room moves every opinion after it. Collect the scores, then discuss the disagreements rather than the candidates: a criterion where two interviewers are four points apart is either badly defined or evidence that they saw genuinely different things, and both are worth ten minutes. The assessment scoring sheet on this site takes the criteria, their weights and each candidate's scores and returns the weighted result and the unweighted one side by side.
Read the weighted and unweighted answers together
The pair is the useful output, not the winner. When the weighted and unweighted rankings agree, the decision is robust and you can stop. When they disagree, the weighting decided it, and that is worth knowing out loud before an offer goes out: it means the candidate is ahead because of what you said mattered, not because they scored higher overall. Either answer is defensible. What is not defensible is a ranking whose sensitivity to the weights nobody ever looked at.
What assessment does not settle
It does not settle whether you should be hiring at all, and it does not price the hire. Those are different figures with their own sheets here: cost per hire for what the process actually costs, time to hire for where the days go, pipeline conversion for how many applicants a hire needs (on the worked example, 50), and the agency fee sheet for whether paying a fee beats filling it in house. Assessment ranks the people in front of you; it does not tell you whether enough of them turned up.
Questions people ask about candidate assessment
What should a candidate assessment measure?
Only the things that decide whether someone can do the job, three to five of them, each with a written description of what a low and a high score look like. Anything you would not reject a candidate over does not belong on the sheet, because it dilutes the criteria that do.
How many interviewers should score?
Enough that no single opinion carries the decision, and few enough that scheduling does not add a week to the process. Two or three is normal on a small desk. What matters more than the number is that they score independently before they talk, so the first opinion in the room does not set the rest.
Should every criterion be weighted the same?
Rarely, because the parts of a job are rarely equally important. Weighting is how you say so before you meet anyone. Run the weighted and unweighted rankings side by side afterwards: if they agree, the weights did not decide it, and if they disagree, they did and you should say so out loud.
Does an assessment score justify a rejection?
It records the basis for the decision, which is worth having, but it is not a substitute for the decision. Use the ranking to order the shortlist and to see how close the top of it is; a two point gap on a hundred point scale is a tie honestly reported, not a winner.