InterviewKata
Behavioural interviews · Module 01 of 12 · Level 1: How it works

How behavioural interviews are scored

Behavioural answers feel like conversation, but they're scored like an exam. The interviewer has a short list of competencies, a scale for each, and a sheet to fill with evidence. Knowing what goes on that sheet changes how you answer every question in this track.

1. The idea

A judge at a diving competition doesn't score how nice the diver seems. They watch for specific things (the take-off, the rotations, the entry) and score each against a published scale. A behavioural interviewer is the same: a short list of things to look for, a scale for each, and a sheet to write the evidence on.
  1. Each question targets a competency. "Tell me about a disagreement" tests conflict; "a time you missed a deadline" tests ownership and dealing with failure. The interviewer usually has two or three competencies to cover in the round.
  2. They collect evidence, not impressions. While you talk, they write down what you did, why, and what happened. Adjectives ("I'm very collaborative") produce no evidence.
  3. Evidence is scored against a rubric for each competency, with a bar that rises with the level: a senior engineer's story about conflict should span more people and higher stakes than a new graduate's.
  4. Others decide from the notes. In a debrief, and at some companies a hiring committee, people who never heard you read what was written. The answer you gave matters only as far as it was written down.

The signals that turn up in almost every rubric:

2. How it works

The interviewer's sheet fills up one line at a time. Watch which parts of one answer become evidence, and which leave no trace.

Most answers lose points the same wayNot by saying anything wrong, but by spending ninety seconds on context and "we", so the sheet has two lines when it needed six.

3. Try it yourself

Note or noise?

Here are eight sentences from different answers. For each, decide: would an interviewer write it down as evidence? Then check.

4. What interviewers look for

Every competency is scored on the same few signals. The bar for each rises with the level you're interviewing for.

SignalWeakGoodStrong
Ownership"We decided…", "the team built…""I proposed…", with one clear action"I noticed nobody owned it, so I…": acted beyond the job, without being asked
Specificity"I usually…", "I always try to…"One real situation, with some detailNames of systems, dates, numbers; details that get richer under follow-ups
Impact"It went well"A result in wordsA measured result, and why it mattered to the customer or the business
ScopeA task assigned to youA feature or project you ledA problem across teams, or one that changed how the team works
JudgementNo reason given for decisionsReasons for the main choiceOptions weighed, trade-off named, and what you'd do differently

5. A model answer

"Tell me about a time you improved a process on your team." An invented answer, shown part by part. Read it for the shape, then build your own from a true story.

SituationOur team released the mobile app every two weeks, and each release took a full day of manual testing by two engineers. Last year, three releases slipped because that testing found bugs too late to fix.

TaskNobody owned the release process: it rotated. I wasn't asked to fix it, but I'd run the last two releases and could see where the day went.

ActionI timed each step of the next release myself and found that 60% of the day was re-testing the same login and payment flows. I wrote automated tests for those two flows first, because they caught most of the late bugs. I ran them in CI on every merge, and showed the team a week of results before asking anyone to change how they worked.

ResultManual testing dropped from a day to about two hours, and in the next six months no release slipped. Two other teams copied the setup.

LearningI'd involve QA earlier. I built the first version alone, and they spotted two flows I'd missed only when I showed them.

Why it scoresEvery sentence leaves a line on the interviewer's sheet: a measured problem, an action you took without being asked, a reason for the order of work, numbers in the result, and a lesson. It's about ninety seconds long, which leaves time for follow-ups.

6. Remember it

Memory hook: "Evidence, not adjectives"Interviewers write down what you did and what happened. Describing yourself leaves the page blank.
How this hook works: what “Evidence, not adjectives” means

What the slogan means, phrase by phrase

“Evidence”
Actions you took, decisions with reasons, results with numbers: things a reader can check against a rubric.
“not adjectives”
"Passionate", "collaborative", "detail-oriented": claims anyone can make, so they earn nothing.

Why this picture fits: A courtroom works the same way. The jury can only weigh what's in the record; a lawyer saying "my client is honest" adds nothing to it.

The story → the concept

The courtroomThe interview
The witnessYou
The court reporterThe interviewer taking notes
Exhibits and testimonySpecific actions, numbers, decisions
"My client is a good person""I'm a team player"
The jury reading the recordThe debrief or hiring committee

Watch it on an example

  1. You say: "I'm really good at handling conflict." The sheet stays empty.
  2. You say: "I asked the two leads for a 20-minute call, and we agreed to run both designs on one week of traffic." One line of evidence.
  3. You add: "The A/B test settled it; we shipped the cheaper design and saved about $4,000 a month." A second line, with a number.

The trap it saves you from

Long, pleasant answers that feel good in the room and score low on paper.

It still works when…

Your interviewer is warm and chatty. The notes still decide, and someone else still reads them.

7. Common mistakes

Describing yourself"I'm a natural leader" is a claim, not evidence. Tell the story and let the interviewer reach the conclusion.
Too much contextIf the situation takes a minute, the action gets thirty seconds. Keep situation and task to about a fifth of the answer.
"We" all the way throughThe interviewer can't score the team. Say what you did, and credit others by name where it's true.
No resultStories that end at "so we shipped it" leave the most important line blank. Say what changed, with a number if you can.
Hypotheticals"I would…" answers a different question. If you have no example, say so and give the closest real one.

8. Practice exercise (25 minutes)

  1. Find the competencies (5 min). Pick a job you'd apply for. From its description, list the four or five behaviours it obviously cares about.
  2. Record one answer (5 min). Answer "Tell me about a time you improved something on your team" out loud, recorded, without notes.
  3. Be the note-taker (10 min). Play it back and write only the lines an interviewer could use as evidence: actions, reasons, numbers.
  4. Count and fix (5 min). Fewer than four lines? Find where the time went (usually context and "we"), and say it again.

9. Self-check quiz

What is a behavioural interviewer mainly doing while you answer?

They take notes against a short list of competencies, which are scored later.

Which sentence gives the interviewer evidence?

It's a specific action you took; the others are claims.

How does the bar change at more senior levels?

The rubric is the same shape, but scope, influence and judgement must grow with the level.

Why do numbers in the result matter so much?

A number is easy to note, hard to dispute, and shows the work mattered.

Roughly how much of an answer should go on situation and task?

Spend most of the time on what you did and what happened.

10. Go deeper

ResourceTimeWhy
How we hire (Amazon)10 minAmazon's own description of its loop, Bar Raisers and how to prepare.
re:Work (Google)15 minGoogle's public guides on hiring, including structured interviewing: why interviewers use set questions and rubrics.
Module 02: The STAR methodNextThe shape that makes every sentence count.