Interview scoring rubric: a free template with behavioural anchors

The template: copy it

Use one row per competency. Keep the question and anchors identical for every candidate, and add a column each for the score and the evidence behind it.

Under the table, add an overall recommendation (no hire, hire or strong hire), the interviewer’s name, the date, and whether it was scored before the debrief.

What are behavioural anchors?

Behavioural anchorsare concrete descriptions of what a candidate says or does at each point on a rating scale. Instead of “3 = meets expectations”, an anchor says “3 = gives a specific example with clear actions and a result, but doesn’t explain why they chose that approach.” A rating scale built this way is called a behaviourally anchored rating scale.

Anchors give interviewers a shared picture of each score. Without them, one person’s 4 is another’s 3, and the debrief turns into a debate about the scale rather than the candidate. Structured interview guidance from public bodies, such as the US Office of Personnel Management’s structured interview guide, recommends defined rating scales for this reason.

What does a good one-to-five interview rating scale look like?

You don’t have to anchor all five points. Writing anchors for 1, 3 and 5 and letting 2 and 4 sit between them is common and easier to maintain.

The STAR method (situation, task, action, result) is a useful lens here: weaker answers usually miss the action or the result.

Worked examples for three roles

Software engineer: debugging under pressure.“Tell me about a production issue you helped resolve. What did you do first?”

• 1: talks about the outage in general terms; cannot say what they personally did.
• 3: describes the steps they took to find and fix the cause, and the outcome.
• 5: explains how they narrowed down the cause, what they ruled out and why, how they communicated during the incident, and what they changed afterwards so it would not recur.

Customer support: handling an upset customer.“Describe a time a customer was angry about something you couldn’t fix straight away.”

• 1: blames the customer or the company; no example of what they did.
• 3: explains how they acknowledged the problem, set expectations and followed up.
• 5:as 3, plus how they adapted their tone to this customer, what they escalated and why, and how the customer’s view changed by the end.

Account executive: recovering a stalled deal.“Tell me about a deal that went quiet. How did you get it moving again, or decide to walk away?”

• 1:describes “following up a lot”; no insight into why the deal stalled.
• 3: identifies why it stalled and the specific steps they took with the buyer.
• 5: as 3, plus how they tested whether the deal was real, who else they brought in, and the result, including when walking away was the right call.

If you cannot write the evidence for a score, the score is a guess.

How do you score a structured interview consistently?

1. Score independently, straight after the interview. Fill in the rubric before you talk to anyone else on the panel.
2. Write evidence for every score. A short quote or paraphrase of what the candidate said. If you cannot write one, the score is a guess.
3. Score each competency on its own. Do not let a great answer on one raise the others.
4. Run the debrief in a fixed order. Have each interviewer share scores and evidence before any open discussion, and let the most junior person go first so seniority does not set the tone.
5. Decide against the bar, not against other candidates. Agree before interviews start what score counts as a hire.

Consistency between interviewers is called inter-rater reliability. Anchors, evidence notes and independent scoring are the simplest ways to improve it.

How many competencies should an interview rubric have?

Three to five per interview is a practical limit. With more, interviewers rush questions and the scores get shallow. If a role needs more competencies, split them across interviewers so each covers a few in depth.

What are the most common rubric mistakes?

• Vague anchors.“Good communication” at level 4 tells nobody anything. Describe what the candidate actually says.
• Too many competencies. Ten competencies in 45 minutes means none are tested well.
• The halo effect. One impressive answer, school or employer colours every score. Independent scoring per competency is the fix.
• Changing questions between candidates. Scores only compare if everyone got the same questions.
• Scoring after the debrief. Once you have heard others, your own view shifts. Score first.

How do you use a rubric with AI interviews?

The same rules apply: one interview plan and one rubric for every candidate, with each score tied to evidence. In EmployLabs’ AI interviewer, Naira, every candidate gets the same plan and rubric, and each score links to a quote from the transcript, so you can check the reasoning yourself instead of trusting a number. You can see what that looks like in a sample report, read how every score comes with its evidence, and learn how to read an AI interview report. AI scoring supports a human decision; it does not replace one. Use the report to decide what to probe in your own interview.