How to Build a Structured Interview Scorecard, Step by Step

To build a structured interview scorecard, choose four to six criteria that predict success in the role, define what each point on your rating scale means in plain words, attach one or two questions to each criterion, and have every interviewer score on their own before anyone discusses the candidate. The scorecard is the same for every candidate in the role, which is what makes the comparison fair and the final decision easier to explain.

Why a scorecard beats a general impression

Without a scorecard, each interviewer remembers a different conversation and weighs different things. One person liked the candidate's energy, another noticed a gap in technical depth, and the debrief turns into a contest between impressions. A scorecard forces everyone to judge the same criteria with the same scale. It also leaves a record you can look back on if a decision is questioned. It does not remove judgement. It points judgement at the right things. A useful test of your scorecard is to ask whether two interviewers who saw the same answer would give it the same score. If they would not, the criterion or the scale is too vague. Fix the wording before you use it, not after the first disagreement in a debrief. The same structure helps new interviewers. Someone who has never interviewed for the role can read the criteria, the scale and the sample questions and run a decent interview on their first try. It also protects the candidate: people are judged on what they showed against a known standard, not on how well they got on with one person. Keep scorecards for a role in one place so that a later search for a similar job starts from what you learned, and update them when a hire shows that a criterion mattered more or less than you thought. Do not turn the scorecard into a formula that makes the decision for you. A total score is a summary and a prompt for discussion. A candidate with a lower total but a clear strength on the one criterion that matters most may still be the right choice, as long as the team can say why.

Start from the job, not from a template

List what a person needs to do well in the first year. Read the job description and talk to the hiring manager. Turn each real demand into a criterion, such as writing clear customer replies, managing a budget, or debugging a production issue. Generic criteria like "culture fit" or "energy" are hard to score and invite bias, so replace them with the specific behaviour you want to see. Keep the list short. Four to six criteria can be covered properly in an hour. Ten cannot. A practical way to find criteria is to ask the hiring manager to describe the best person who ever did this job and the one who struggled. The differences between the two are usually your criteria. Avoid criteria that only restate the job title, and avoid anything you cannot observe in an hour, such as long-term loyalty. If a trait matters but cannot be seen in an interview, test it another way, for example with a short work sample or a reference call.

Split the criteria between interviewers

Not every interviewer needs to cover every criterion. Give each stage two or three criteria so that nobody repeats the same questions and the candidate is not asked the same thing four times. Make sure each criterion is owned by at least one person, and the most important ones by two, so a single bad day does not decide the result. Share the plan with the interviewers a day or two before the first interview, with the criteria they own, the questions and the scale. A ten-minute briefing is enough. Without it, interviewers improvise, and the point of structure is lost.

Define the scale in words

A number alone means different things to different people. Write a short description for each point. For a five-point scale, for example, a 1 might mean no evidence of the skill, a 3 might mean a solid, job-ready example, and a 5 might mean a strong example that goes beyond what the role needs. Use the same definitions for every candidate and every interviewer. If you prefer fewer levels, a four-point scale removes the safe middle and forces a lean one way or the other. Either is fine if the words are written down. Anchors do not need to be long. One sentence per level is enough, and examples help: for a criterion such as stakeholder communication, a low score might describe someone who could not give a specific example, and a high score might describe someone who explained how they handled a disagreement and what changed afterwards. Share the definitions with interviewers before the first interview, and take ten minutes to talk through them so that the team uses the words the same way.

Attach questions and a look-for

For each criterion, write one or two questions that ask for real past behaviour or a real task, plus a line on what a good answer contains. This is the same idea as a structured interview: every candidate gets the same core questions, in the same order, so differences in the scores come from differences in the answers. Interviewers can still ask follow-ups, but they should record the core answer before they wander.

Score on your own, then compare, and keep notes about evidence

Score on your own, then compare. Ask each interviewer to write notes and scores straight after the interview, before they talk to anyone else. Early conversation pulls people towards the loudest opinion. When scores are in, the debrief can start with where they differ, which is where the useful discussion is. An interviewer who scored a 2 and one who scored a 4 on the same criterion should each point to what they heard. Keep the notes about evidence. Record what the candidate said or did, not how you felt. "Described rebuilding the onboarding flow and named the trade-offs" is evidence. "Seemed sharp" is not. Leave out anything unrelated to the job, such as age, family plans, accent or background. In many countries, interview notes can be requested by the candidate or reviewed in a dispute, so write only what you would be comfortable defending. Allow a short time after the interview for scoring, ideally the same day, because memory fades quickly and later notes drift toward the final impression. If an interviewer wants to change a score after the debrief, ask them to say why in writing, so that the record shows the reason and not only the new number. Review the pattern after a few hires: if one interviewer always scores high or always scores low, a short calibration chat helps, and it is about the scoring habit, not the person.

Rules differ by country

Laws on discrimination, data access and record keeping differ widely. In some places candidates have a right to see the personal data held about them, including interview notes. In others, particular questions about family, health or beliefs are restricted. Check the rules where you hire and where the candidate lives.

In Rezumz

On a job in Rezumz, the interview scorecard panel lists what every interviewer is asked to rate, from 1 to 5. By default it is the job's must-have skills plus communication, and you can edit the list (up to 10 items, each with an optional note on what a strong answer shows). Every new interview link starts from it, so each interviewer gets their own link and form, and scores can be compared across interviewers. Submitted feedback is locked, and a changed view is saved as a new version with a reason. Rezumz does not decide who to hire. The scorecard gives your team a shared structure, and people make the decision.