How to score interview answers using a rubric

To score interview answers using a rubric, decide the competencies before anyone walks in, write down what a weak, solid and strong answer looks like for each question, score every candidate against those anchors, and only then compare notes. If you recruit for clients, knowing how they score is also the fastest way to prep candidates who score well.

Last reviewed August 2026.

Why gut-feel interviews cost you placements

You've sat in the debrief. The hiring manager says your candidate "didn't quite land" or wasn't a "culture fit", and when you push for specifics there aren't any. What usually happened is that the interviewer leaned on intuition, and intuition tends to favour people who talk like us, look like us or went to the same university. That's unfair on the candidate, and it's expensive for everyone: a placement lost on vibes, a role still open, a client no closer to a hire.

Unstructured interviews are poor at predicting how someone will actually perform, because without structure you're asking different questions of different people and then comparing apples with oranges. An interview rubric is the guardrail. It forces the panel to decide what success looks like before they meet anyone, and it shifts the conversation from "do I like this person?" to "did this person show us the behaviours the role needs?"

For a recruiter, that shift matters twice. It gives your strong candidates a fair hearing, and it tells you exactly what to prepare them for.

Step 1: define the competencies

Before anyone can score an answer, the panel needs to know what it's looking for. Most job descriptions are a list of years of experience and vague traits like "hardworking". A rubric turns those into measurable competencies, and the way to get there is to think about the work the role actually involves.

Compono's research into high-performing teams describes eight ways of working that every team needs covered: Doer, Auditor, Helper, Advisor, Pioneer, Campaigner, Evaluator and Coordinator. Does this role need someone who finishes what they start (Doer), someone who spots the risk everyone else missed (Evaluator), someone who keeps the plan on track (Coordinator) or someone who keeps the team together under pressure (Helper)? Once you know which of these the role leans on, the competencies write themselves. If you want the framework in front of you while you brief a client, the free four-question work personality assessment is a quick way to see the eight types.

Aim for four or five competencies per role. Write a one-line definition for each so every interviewer means the same thing. "Leadership" could mean directive command or democratic collaboration; if the panel hasn't agreed which, the goalposts move mid-process and your candidate pays for it.

Step 2: anchor every score to observable behaviour

A standard interview scoring rubric uses a 1 to 5 scale, and the numbers are meaningless on their own. A "3" shouldn't mean "average". It should mean "gave a specific example that shows the basic requirement, with some guidance needed". A "5" should mean "specific, high-impact example, explained the reasoning, considered the effect on the wider team". A "1" is vague, generic, or shows the competency isn't there.

The most common rubric mistake is levels like "poor, good, excellent" with nothing underneath them. The fix is to replace every label with what an interviewer could actually observe in the answer. Here's a sample row.

Sample interview rubric row

Do this for every question and interviewers can disagree about a candidate and actually resolve it, because they're pointing at evidence rather than a feeling.

Step 3: calibrate the panel

A rubric is only as good as the people using it. If one interviewer is a hard marker and another hands out 5s like lollies, the data is skewed and your candidate's fate depends on who was in the room. Before interviews start, have the panel score a hypothetical answer together and talk through any gap. That conversation is where the rubric actually gets agreed.

During the interview, the interviewer's job is to listen and take notes on what was actually said: the verbs, the decisions, the outcomes. Scoring while the candidate is still talking invites the halo effect, where one great early answer papers over weak later ones. Score in silence after the candidate leaves, one question at a time.

The other bias worth naming is the mirror. Interviewers naturally score higher the answers that sound like they'd have given themselves. An Evaluator on the panel warms to crisp, data-first answers; a Campaigner warms to energy and story. Knowing your own work personality doesn't remove that pull, but it lets you notice it, and it's why the best hiring panels mix types on purpose. The goal is a team that covers the ways of working it's currently missing.

Step 4: use the scores to make the call

With every candidate scored against the same anchors, the decision meeting changes character. Instead of debating vibes, the panel debates evidence: "Candidate A scored a 5 on problem-solving and a 2 on collaboration. Candidate B was a solid 4 across the board." That doesn't mean the top total always wins. Some competencies are non-negotiable, and if the role needs Auditor-level attention to detail and someone scored a 1 on it, charm elsewhere doesn't rescue them. The rubric gives the panel licence to say no to someone they liked, and yes to someone who was quieter but demonstrably better.

It also gives you something recruiters rarely have: specific, defensible feedback for the candidates who didn't get it. "You were strong on delivery and light on examples of influencing without authority" is useful to a person. "They went with someone more experienced" is not, and candidates remember which one you gave them.

What this means for candidate prep

Most recruiters skip this part. Once you understand how a client scores answers, you know exactly how to prepare candidates to score well. Tell them the competencies (you don't need to hand over the sheet). Teach them to answer with a specific situation, the action they took, the result and, crucially, the reasoning behind it, because that reasoning is what separates a 3 from a 5 on almost every rubric. Then get them to say it out loud, more than once, before the real thing.

That last step is where most prep breaks down. A phone briefing and a link to a YouTube video doesn't move a candidate from a 3 to a 5. Practice does, and you can't personally run mock interviews for every candidate on every brief. That's the gap candidate prep exists to close: every candidate you put forward gets AI interview practice, role play and scored feedback as part of your placement service, and you see who's actually ready.

Can you auto-score interview responses?

For practice, yes, and it's useful. An AI coach can score a candidate's answer against behavioural anchors and tell them what was missing, which is how Hey Compono's role play works for the candidates recruiters put through Recruiter Hub. It's private to the candidate by default, and they can choose to share a standout result with you.

For the hiring decision itself, keep humans doing the scoring. Automated scoring of live interviews raises fairness and legal questions that vary by jurisdiction, and it removes the calibration conversation that makes a rubric work. Use automation to help candidates practise and to keep interviewers consistent. Let people make the call.

Related reading: How to prep a candidate who is nervous for an interview · How does AI interview roleplay work?

Frequently asked questions

What is an interview rubric?

An interview rubric is a structured scoring tool that measures every candidate's answers against the same predefined criteria. It usually pairs a numerical scale with a written description of what a weak, solid and strong answer looks like for each question.

How do interview rubrics reduce hiring bias?

Rubrics make interviewers focus on specific, observable behaviours and evidence rather than a feeling. Because the panel agrees what a good answer looks like before the interview, a candidate's accent, background or resemblance to the interviewer has less room to sway the score.

Should you share the rubric with the candidate?

You don't need to hand over the scoring sheet, but telling candidates which competencies matter is fair and useful. It lets them choose relevant examples, which gives the panel better evidence to score against. Recruiters should brief candidates on the competencies as standard.

What is the fix when a rubric uses vague levels like good and excellent?

Replace each label with an observable behavioural anchor: what a candidate at that level actually says or demonstrates. For example, a 3 gives a specific situation, action and result; a 5 also explains the reasoning and the impact on others. If two interviewers could read a level differently, it needs rewriting.

What happens if interviewers give very different scores?

Large gaps usually mean the competency wasn't defined clearly enough or the interviewers were listening for different things. Hold a short calibration meeting, compare the evidence each person wrote down against the anchors, and agree the score. Then fix the rubric so the gap doesn't recur.