Most interview scorecard templates I've found are a list of competencies with a five-point scale next to each one: communication, problem solving, leadership, culture add. They look rigorous and score nothing, because nobody can test "leadership" in 45 minutes, so the column fills with 3s and the debrief goes back to impressions. The rows are the problem, not the scale. This is for recruiters who run interview loops across several functions and need one scorecard shape that hiring managers in sales, CS, engineering and ops will all fill in. You'll leave with a template, four rules for filling it, and worked rows for four roles.
My bias, up front. Skillvee is a 60-minute "day at work" simulation that replaces the recruiter phone screen and the first interview round. The candidate solves a real challenge, talks to AI teammates to get the information they need, and defends their decisions on a call, while the screen is recorded. So I think about scorecards as what turns an hour of observed work into a decision, and I've watched plenty fail at it. The template below works with interviews, work samples, or both.
What an interview scorecard is for
An interview scorecard is the form each interviewer fills in after their conversation with a candidate, and its only job is to make several people's reads comparable.
That's narrower than most templates assume. A scorecard is the few things you decided to test, a score for each, and the evidence that earned it, which makes it smaller than the job description and the interviewer's notes. If two interviewers can't put their sheets side by side and see where they disagree, the sheet has failed, however thorough it looks.
The rows are where that comparison lives or dies. "Problem solving" means something different to every interviewer who reads it. "Finds the at-risk renewal in the account data" means one thing, and two people who scored it differently heard different things, which is what you want the interview debrief to find.
The one-row-per-task scorecard: four rules
I call this the one-row-per-task scorecard. Four rules, and the first one does most of the work.
- A row is a task, not a trait. Every scored row describes something the person will do in their first 90 days, phrased so you'd recognise it happening. If you ran the hiring manager intake meeting the way I laid it out, these are the signals that survived the task test, and the rows are already written. If you didn't, take each competency the manager wants and ask where they'd see it in the first three months. What can't be tied to a task is a preference and stays off the sheet.
- Knockouts sit above the rows and never get a score. Work authorization, location, compensation range, a licence, a hard must-have. They're yes or no, mostly checked on the phone screen, and giving them a number lets a strong candidate's 4s hide a missing requirement.
- A score only exists next to a quote. The scale is 1-4, the same one the recruiter phone screen template uses, so the screen and the panel read the same way. Four takes away the middle. Next to every number goes what the candidate said or did, close to their words. No quote, no score.
- Every row says where the evidence came from. Which interview, or the work sample. A row that nobody tested stays blank. A blank is honest and a 2 "because I didn't see it" isn't, and once three sheets land in a debrief, a blank reads as a gap in the interview plan instead of a weak candidate.
Here's what the four rules produce for an account executive, scored by the hiring manager after a role play.
| Scored row (the task) | What a 4 looks like | Score | Quote or action | Seen in |
|---|---|---|---|---|
| Runs discovery | Asks who else signs off and what happens if they do nothing, before any feature talk | 4 | "Before I show you anything, who else would need to be comfortable with this?" | Role play |
| Handles a pricing objection | Asks what the number is being compared to instead of discounting | 2 | Offered 15% off within a minute of hearing "too expensive" | Role play |
| Sets a real next step | Proposes a dated meeting with a named person and a purpose | 3 | "Can we get your ops lead on a call Thursday?" No agenda offered | Role play |
| Written follow-up | Short, names the objection and what changes it | Blank | Not tested in this interview |
Knockouts, above the rows and not scored: can sell in US time zones, base within range. Both confirmed on the screen.
Three rows scored, one blank, and every number has words next to it. A week later, anyone reading this knows what the 2 was.
The interview scorecard template
Copy this per interviewer, per candidate. The hiring manager fills in the left column before the first interview and the panel never edits it.
INTERVIEW SCORECARD
Candidate: [name] Role: [role] Interviewer: [name]
Interview or exercise: [which one] Date: [date]
KNOCKOUTS (yes / no, not scored)
[ ] [requirement 1] [ ] [requirement 2] [ ] [requirement 3]
SCORED ROWS (1-4, no score without a quote)
Row 1: [task]
A 4 looks like: [one sentence agreed at intake]
Score: [ ] Quote or action: [their words, or what they did]
Row 2: [task]
A 4 looks like:
Score: [ ] Quote or action:
Row 3: [task]
A 4 looks like:
Score: [ ] Quote or action:
NOT TESTED: [rows you didn't get to, left blank above]
ADVANCE RULE (written at intake, same for every candidate):
[for example: no row below 2 after the debrief, and at least two rows at 3 or better]
SUBMIT BEFORE READING ANYONE ELSE'S.
The "a 4 looks like" line is the one most templates skip, and it's what makes scores comparable. If the manager and the peer interviewer carry different pictures of a 4, you'll find out in the debrief, at the worst time. Write it once, at intake, in a sentence you could read aloud.
The same scorecard for four roles
The shape doesn't change across functions. Only the task in each row does. One row per role, so you can see nothing in the template is sales-specific.
| Role | A scored row (the task) | What a 4 looks like |
|---|---|---|
| Customer success manager | Finds the renewal that's quietly at risk, in a case study | Asks who the executive sponsor is before presenting a plan, when the usage data looks fine |
| Backend engineer | Picks up a change in an unfamiliar codebase, in a work sample | Reads the existing tests and asks what the caller expects before touching the function |
| Program manager | Runs a status meeting with a slipped dependency, in a role play | Names the slip, the owner and the new date in the first five minutes, without being asked |
| Account executive | Handles a pricing objection, in a role play | Asks what the price is being compared to instead of reaching for a discount |
The CSM row is the buried-signal case from my customer success interview case study, and the engineer row is what a good work sample tests when the candidate has an AI assistant open: whether they found out what the code was supposed to do. The columns stay identical, so a recruiter with six open roles keeps one template and a library of rows.
Filling in the quote column
The quote column is where scorecards go wrong in practice.
Write during the interview, not after. Two days later, most people remember their verdict and not the sentence that produced it. One line per row is enough.
When an answer sounds polished and general, don't score it yet. Ask for a second example from a different company, then go a level down. That's the second-example probe from the phone screen template, and the quote you write down is whatever came out at the second layer.
Observed actions count the same as quotes. "Opened the test file before the source" is evidence. "Seemed methodical" goes in the margin, if anywhere.
Where this gets hard
Hiring managers will push for trait rows. "I need to know if they're strategic." You can't score strategic, but you can score what they'd do in the first QBR they run, and that's the row. Most managers accept this once they've named the task themselves. A few won't, and then you're scoring a word, and the debrief will be about the word.
Interviewers hate blanks. Left alone, a row nobody tested gets a 2 or a 3 so the sheet looks complete. The "not tested" line makes the blank deliberate, and say so at kickoff.
Weights are tempting and I'd mostly skip them. Once rows are tasks, the manager can say which two matter most and the advance rule can lean on those. A weighted average hides the row you needed to argue about. If you do weight, do it at intake, before anyone is scored, and write the weights on the sheet.
And a scorecard is only as good as the exercise it scores. Four rows about real tasks still produce thin evidence if the interview is 45 minutes of questions about the resume. That's a big part of why Skillvee gives the candidate an hour of actual work and records it: the row gets filled with what the candidate did, and a disagreement in the debrief can be settled by replaying it.
What to do with the scorecards after the loop
Collect every sheet before anyone reads a colleague's, put the scores in a grid, and run the debrief on the rows where the panel sits two or more points apart. The hiring manager decides against the advance rule on the sheet, and the record goes back to the req.
After three or four candidates on the same role, look at which rows keep coming back blank. That's a task you said mattered and nobody is testing. Add an exercise for it, or admit it's a preference and drop the row. If the row that keeps going untested is how someone works when nobody's asking them questions, that's the part an interview can't reach. The pricing page lists the plans if you want to try a simulation on that one role.