An interview scorecard template for small teams: four competencies, scored anchors, one page
An interview scorecard template with four competencies, 1–4 anchors and a copyable form, plus the research on why structured interviews beat gut feel at a.
On this page
An interview scorecard is a one-page form that every interviewer fills in independently, before talking to anyone else, scoring the candidate on the same four to six competencies against written anchors for each score. It turns "I liked her" into "3 out of 4 on problem-solving, because she worked through the caching question without hints but missed the invalidation case." The template below has four competencies, anchors for scores 1 to 4, and a copyable form, and it takes about ten minutes per interview to use.
We are a company of under fifteen people and we adopted scorecards after a hire that three of us had rated "great" in conversation and none of us could explain afterwards. This post is the version we use now and the evidence for why it is worth the ten minutes.
Why unstructured interviews fail at a five-person company
At a large company, a bad interview is one data point among many. At five people, the interview usually is the process, and an unstructured one has three problems that small teams feel immediately.
Every interviewer measures something different. Without agreed competencies, one person probes technical depth, another talks about culture, a third spends the hour selling the company. The debrief compares impressions that were never of the same thing.
The first opinion anchors the rest. When the founder says "I loved him" in the first minute of the debrief, the two people who had doubts soften them. Independent written scores, submitted before the debrief, are the only reliable fix.
You cannot learn from it. Six months on, you want to know which interview signals predicted performance. Without a score and a reason per competency, there is nothing to look back at.
The research is older than most hiring software. Schmidt and Hunter's 1998 meta-analysis in Psychological Bulletin reported a validity of .51 for structured interviews against .38 for unstructured ones in predicting job performance. Sackett, Zhang, Berry and Lievens revisited those estimates in 2022 in the Journal of Applied Psychology, correcting what they argued was over-correction for range restriction in the earlier work; their revised figures put structured interviews at about .42 and unstructured interviews at about .19. The absolute numbers moved, the ranking did not: in both analyses, adding structure roughly doubles how much an interview tells you.
The four competencies
Four is the right number for a small team. More than that and interviewers start scoring from memory rather than evidence; fewer and the role-specific skill crowds out everything else. Adjust the names to the role, keep the shape.
| Competency | What it means here | How you get evidence |
|---|---|---|
| Role skill | Can they do the core work of this job at the level the role needs | A work sample or a technical walkthrough of something they built |
| Problem-solving | How they approach an unfamiliar problem: questions, structure, trade-offs | A problem from your actual backlog, simplified; watch the process, not just the answer |
| Communication | Can they explain their thinking so a teammate can act on it, in writing and aloud | Ask them to write up the problem-solving answer in five minutes; read it later |
| Ownership | Do they finish things, and tell people early when they will not | Behavioural questions about a project that went wrong; listen for what they did, not what the team did |
The anchors: what each score means
Anchors are the part most scorecards skip, and the part that makes scores comparable across interviewers. Write them before the first interview, per competency. Here is the set for problem-solving; the full form below has all four.
| Score | Problem-solving anchor |
|---|---|
| 1 | Jumps to a solution without clarifying the problem; cannot adjust when a constraint is added |
| 2 | Asks some clarifying questions; reaches a workable answer with significant hints |
| 3 | Structures the problem, names trade-offs, reaches a sound answer with minimal hints; misses one edge case |
| 4 | Does all of the above unprompted, identifies the edge cases themselves, and proposes how they would test it |
A 4 should be rare. If every interviewer gives 4s, the anchors are too soft.
The scorecard, ready to copy
One per interviewer per interview. Fill it in within an hour of the interview and before any conversation with other interviewers. Labels are in bold so the form stays readable in a code block.
**Candidate:** [name] **Role:** [title]
**Interviewer:** [name] **Stage:** [screen / work sample / team / final]
**Date:** [date] **Duration:** [minutes]
**1. Role skill** Score: [1 / 2 / 3 / 4]
Anchors: 1 = cannot do the core task; 2 = does it with help;
3 = does it independently to our standard; 4 = does it and improves on
our standard.
Evidence (what they did or said, not your impression):
-
-
**2. Problem-solving** Score: [1 / 2 / 3 / 4]
Anchors: 1 = solution before problem; 2 = workable with hints;
3 = structured, sound, one miss; 4 = unprompted, edge cases, test plan.
Evidence:
-
-
**3. Communication** Score: [1 / 2 / 3 / 4]
Anchors: 1 = hard to follow; 2 = clear when asked to clarify;
3 = clear first time, aloud and in writing; 4 = clear and anticipates
what the listener needs.
Evidence:
-
-
**4. Ownership** Score: [1 / 2 / 3 / 4]
Anchors: 1 = describes outcomes as the team's or circumstances';
2 = owns outcomes but surfaced problems late; 3 = owns outcomes and
raised problems early; 4 = owns outcomes, raised them early and changed
the plan.
Evidence:
-
-
**Concerns** (anything that would make you say no, regardless of scores):
**Recommendation:** [strong no / no / yes / strong yes]
**One line on why:**
Two deliberate omissions. There is no "culture fit" line, because it is where bias hides; if there is a specific behaviour you need, make it a competency with anchors. And there is no overall score computed from the four, because a 1 on role skill is a no regardless of the others.
How to run it, step by step
- Write the competencies and anchors for the role before posting it. The same inputs you used for the job description become the anchors; our first-engineers hiring post covers that sequence.
- Assign competencies to interviews. Two per interview, so each gets twenty minutes of evidence. The work sample covers role skill and problem-solving; the team call covers communication and ownership. Nobody tries to cover all four in one hour.
- Give each interviewer the questions, not just the competencies. The same three questions per competency, asked to every candidate. Follow-ups can vary; the openers cannot.
- Score within the hour, independently. Scores go into the form before anyone speaks to anyone. In Atrix we attach the scorecard as a database row on the applicant's record, one row per interviewer, and the hiring manager does not open the rows until all are in.
- Debrief from the evidence columns, not the scores. Read the evidence aloud for any competency where interviewers differ by two or more. The disagreement is usually about what counts as evidence, and that is the calibration conversation.
- Decide on the pattern. A 1 anywhere is a no. Otherwise, hire when the role-skill and problem-solving scores are 3 or higher from the interviewers who saw them, and nothing in "Concerns" is unresolved.
- Keep the scorecards. At the six-month review, read them again. The competencies where scores predicted performance are the ones to weight next time; the ones that did not are the ones to rewrite.
Calibrating across interviewers
Scorecards drift. One interviewer's 3 is another's 2, and the gap grows unless you check it.
- Score a shared candidate. Once a quarter, two interviewers sit in on the same interview and score independently. Compare the sheets. Differences of one point are normal; two points means the anchor needs rewriting.
- Rewrite anchors with examples from real candidates. "Missed the invalidation case" is a better anchor than "misses one edge case" once you have seen it happen.
- Watch the distribution. If one interviewer never gives a 2 or a 4, their scale has collapsed to pass/fail. Show them the distribution.
- Keep the questions fixed for a hiring round. Changing the opener mid-round means candidates are no longer comparable.
Where the scorecard lives
A scorecard in a personal notebook is as good as no scorecard, because nobody else can read it at the debrief or six months later. The options that work: a shared document per candidate with a section per interviewer, a form tool that writes to a spreadsheet, or a record on the applicant inside whatever tracks your pipeline.
In Atrix, the hiring pipeline holds each applicant, and a page database attached to the applicant holds one scorecard row per interviewer, with the competencies as columns so the debrief view is a table. The docs and whiteboards page covers page databases; the hiring feature page covers the pipeline, the careers page and the hire-to-teammate step. What Atrix does not do: it does not score candidates, screen CVs or rank applicants with AI. The scores are yours, typed by the person who did the interview.
FAQ
What is an interview scorecard?
A one-page form each interviewer completes independently after an interview, scoring the candidate on a fixed set of competencies against written anchors, with evidence for each score and a hire recommendation. It is the main tool of a structured interview.
How many competencies should an interview scorecard have?
Four to six. Four for a small team hiring generalists; up to six for a specialised senior role. Beyond six, interviewers score from memory rather than evidence.
Should interviewers see each other's scorecards before the debrief?
No. Independent scores submitted before any discussion are what make the scorecard work; the first opinion voiced otherwise anchors the rest. Open them together at the start of the debrief.
What scale should a scorecard use?
A 4-point scale with written anchors. An even number removes the neutral middle, and anchors make a 3 mean the same thing to every interviewer. A 10-point scale without anchors is a 10-point guess.
Does a scorecard remove bias from hiring?
It reduces it; it does not remove it. Fixed questions, written anchors and independent scoring are the measures with the strongest evidence behind them. Scoring "culture fit" without anchors reintroduces what the rest of the scorecard removed.
Can Atrix score candidates automatically?
No. Atrix stores scorecards against the applicant in the pipeline and shows them as a table at the debrief. Reviewing and scoring candidates is done by your interviewers, not by AI.
Next step
Copy the form above into one page per candidate, write anchors for your role before the first interview, and have every interviewer score independently. If your pipeline is in Atrix, the hiring feature page shows where the scorecard attaches; the applicant tracking glossary entry explains the stages the pipeline uses.
Sources
Schmidt and Hunter, "The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings", Psychological Bulletin 124(2), 1998; Sackett, Zhang, Berry and Lievens, "Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range", Journal of Applied Psychology 107(11), 2022.