Interview Scorecard: Template, Examples, and the Six Parts That Make One Work (2026)

Two interviewers spend 45 minutes each with the same candidate. One writes “strong yes, excellent communicator, would hire.” The other writes “no, vague, could not get a straight answer.” Neither of them is lying, and neither is careless. They reached opposite conclusions because nobody had written down what the interview was supposed to measure, so each of them measured something reasonable and different. The document that prevents this has a name, and most teams build it at the wrong end of the process.

What is an interview scorecard?

An interview scorecard is a short document, agreed before anyone is interviewed, that states the criteria a candidate is being evaluated against, how much each one counts, and what different levels of performance against it look like in observable terms. Interviewers rate the candidate on each criterion and record what they actually heard that justifies the rating.

You will see the same artefact called a candidate scorecard, a hiring scorecard, a recruiting scorecard, an interview evaluation form or an interview scoring sheet. The names are interchangeable. What is not interchangeable is whether the criteria are specific to the role and written down in advance, because that is the only feature doing any work.

Three things a scorecard is not:

• It is not a record of the decision. A form completed after you already know your answer is documentation of a conclusion, not a method for reaching one.
• It is not a generic competency sheet. Communication, attitude and experience rated one to five, reused for every role from warehouse operative to CFO, measures whichever of those things happened to stand out.
• It is not a substitute for judgement. It constrains what your judgement is allowed to be about, then makes the resulting judgements comparable across interviewers.

Why two interviewers disagree about the same candidate

The disagreement at the top of this article is not a people problem. It is what happens when an evaluation has no defined target. Left to itself, an unstructured interview tends to converge on how confident, articulate and familiar a candidate seemed, because those are the qualities that are legible in a conversation whether or not the job depends on them. The consistent finding across decades of selection research is that structured interviews, meaning the same questions rated against defined criteria, predict job performance considerably better than unstructured ones. The scorecard is what makes an interview structured on the rating side, the same way a fixed question set does on the asking side.

There is a subtler failure that scorecards fix, and it survives even in teams that think they are already structured. Two interviewers both give a candidate a 3. That looks like agreement, so nobody examines it. One of them meant “did the job adequately at a smaller company and would need support here”, the other meant “nothing went wrong in the conversation”. Without written anchors, the number is a shared symbol for two unshared judgements, and the average of them means nothing at all.

The six parts of a scorecard that actually works

Most scorecard templates supply two of these six and wonder why the output is thin. Each criterion on the sheet needs all six.

• 1. The criterion.Something the job genuinely requires, phrased narrowly enough to be observed. “Debugging an unfamiliar system” is a criterion. “Technical ability” is a category.
• 2. Why it predicts. One line tying it to the actual work. If you cannot write that line, the criterion is inherited rather than chosen, and it is diluting the weights of the ones that matter.
• 3. The weight. A percentage of the decision, set at kickoff, summing to 100 across the sheet.
• 4. The anchors. A written description of what a 1 looks like and what a 4 looks like, in behaviour you could observe rather than qualities you could attribute.
• 5. The evidence box. What the candidate actually said or did, written before the rating is chosen. This ordering matters more than it sounds.
• 6. The rating. Submitted independently, before the debrief opens.

Part five carries surprising weight. Asking interviewers to write the evidence first, then pick the number, forces the rating to be a conclusion from something specific. Reversing the order lets people pick a number from impression and then hunt for a justification, which is the same process with the audit trail attached.

A free interview scorecard template you can copy

Copy this into a doc, a spreadsheet or your ATS. It takes about twenty minutes per role to fill in, once, and every interview for that role reuses it.

Interview scorecard examples with real anchors

Anchors are the part people struggle to write, so here are worked examples from two very different roles. Notice that every anchor describes something a candidate does, not something they are.

Customer support specialist, criterion: de-escalating an angry customer (weight 30%)

• 1: Describes the customer as the problem, or the example given is one where a policy was simply restated until the customer gave up.
• 4: Gives a specific incident, states what the customer actually wanted underneath the complaint, describes what they conceded and what they held, and can say what the outcome was including whether the customer stayed.

Account executive, criterion: qualifying out of a bad deal (weight 25%)

• 1: Cannot recall walking away from a deal, or frames qualification purely as budget confirmation.
• 4: Names a specific deal they killed while it was still live, the signal that prompted it, roughly what it was worth, and what they did with the time instead.

The test for an anchor is whether two people reading it and hearing the same answer would land on the same number. “Communicates well” fails that test instantly. “Can state what the customer wanted underneath the complaint” passes it. If you want the question side of this, interview questions to ask candidates covers which questions produce evidence against criteria like these, and the screening interview guide covers the shorter version for a first call.

Where the criteria should come from, and why it is not the interview

Here is the structural mistake, and it is nearly universal. Teams build the scorecard when they schedule the first onsite. By then, hundreds of applicants have already been sorted into a shortlist using criteria that were never written down, and the scorecard is being asked to evaluate a group that was selected on a different basis than it is about to be measured on.

The criteria already exist. When somebody reads a resume and decides it is worth a call, they are applying criteria, just privately and inconsistently. Writing the scorecard at the point where you open the requisition, rather than at the point where you start interviewing, means one set of criteria governs the entire process. That single change produces three effects:

• The interview stops re-litigating the screen. If the resume already evidenced eight years of the required stack, the interview should not spend twenty minutes confirming it. Interview for what paper cannot establish.
• The screen tells you what to ask. A candidate who scores well overall but has no evidence at all against criterion three has just written your question list for you. The gaps are the agenda.
• Your funnel becomes measurable. One rubric across stages means selection rates per criterion, per stage, which is the raw material for both recruitment metrics and the fairness checks in adverse impact analysis.

The practical blocker is capacity, not conviction. Applying five weighted criteria to 400 applicants by hand is not something a recruiter carrying six requisitions can do, which is exactly why the shortlist ends up built by skimming the first 40 and the criteria only get written down later. That is the constraint worth removing first, and it is the same argument made in how to screen resumes in bulk and resume screening criteria.

Rating scales: why 1 to 4 beats 1 to 5

Use an even-numbered scale. The midpoint of a five-point scale is where undecided interviewers park candidates they did not really evaluate, and a sheet full of 3s is indistinguishable from no evaluation at all. Removing it forces a direction, which is the smallest possible amount of commitment to ask of someone who just spent an hour with a person.

• 1 Clearly below the bar for this role. You saw evidence against.
• 2 Below the bar, but arguable, or the evidence was thin.
• 3 Clearly meets the bar. Would be fine doing this part of the job.
• 4 Exceeds it, with a specific thing you can point at.

Two rules that matter more than the scale itself. First, allow “no evidence gathered” as a distinct entry, separate from a 2. An interviewer who never got to a criterion must be able to say so, because a guessed 2 is worse than a blank and will be averaged in as though it were a finding. Second, never let an interviewer see anyone else's rating before submitting their own.

Set the weights before you meet anyone

Weights assigned after interviews have started are rationalisation. Everyone can construct a weighting under which their preferred candidate wins, and everyone does it sincerely. Agree the percentages at the kickoff meeting with the hiring manager, in the same session where you agree the job description, and write them down before a single resume is read.

The kickoff question that produces honest weights is not “what matters for this role?”, because the answer is always everything. Ask instead: if you had to hire someone who was a 2 on exactly one of these, which one would it be? That surfaces the real ranking in about ninety seconds, and it usually reveals that a criterion everyone listed as essential is one the team would happily trade. A related discipline, deciding which requirements are genuinely required at all, is the substance of skills-based hiring.

Running a debrief that the scorecard survives

A scorecard filled in honestly and then discussed badly produces the same outcome as no scorecard. The debrief is where the method is usually lost.

Spreadsheet, ATS, or screening tool

The format matters less than the enforcement, but the trade-offs are real:

• Spreadsheet.One tab per candidate, criteria as rows, interviewers as columns, weighted total computed automatically. Free, flexible, and the version most teams should start with. The weakness is that nothing stops someone opening a colleague's column before submitting.
• ATS scorecards. Most modern systems, including Greenhouse-style tools, support per-stage scorecards with required fields and locked submissions, which solves the independence problem properly. The risk is the default template: generic competencies shipped by the vendor, reused across every role, which is the failure mode from the top of this article wearing better software. Replace them per role. Options in our applicant tracking systems list.
• Screening tool for the paper stage. The resume half of the rubric is the part that does not scale by hand, so this is where automation earns its place, with the interview half staying human. The boundary between the two is drawn in recruitment automation.

Seven ways interview scorecards fail

• Scorecard theatre. Completed after the decision, to document it. Detectable by a suspicious absence of low scores for hired candidates.
• Twelve criteria. Weights become noise, and interviewers rate things they gathered no evidence on rather than leave a blank.
• Adjective anchors.“Excellent communicator” is a rating pretending to be a definition. Anchors must describe observable behaviour.
• Averaging over a knockout. A 3.6 weighted total on a candidate who cannot legally do the job. Keep hard requirements out of the weighted section entirely.
• Resume halo. Interviewers who read the resume closely arrive with a prior and score the conversation to match. Distribute criteria across the panel so nobody rates everything.
• Rating the interview, not the candidate. Punishing nerves, an accent or an unpolished answer rewards interview practice, which is not on the scorecard and rarely on the job.
• No connection to the screen. The most expensive one. Rigorous scoring applied to a shortlist that was assembled by skimming means you have measured the wrong forty people precisely, which is the coverage problem described in how to shortlist candidates from resumes.

Do scorecards make hiring fairer, honestly?

Partly, and it is worth being precise about the mechanism rather than claiming more than is true. A scorecard narrows what an interviewer is permitted to consider and requires each rating to be tied to stated evidence, which shrinks the space in which impressions about confidence, background or similarity operate. It does not make anyone neutral. A scorecard whose criteria include “culture fit” with no anchors has done nothing at all except give a gut feeling a number and a paper trail.

The second benefit is more reliable than the first and gets discussed less. Consistent criteria make the process measurable. When every candidate has been rated on the same five things, you can compute selection rates by stage and by group and see where a stage is filtering unevenly, using the four-fifths method in adverse impact. Free-text interview notes cannot be analysed at all, so a process without scorecards is not neutral, it is merely unexamined. It is also the difference between a defensible hiring decision and one that rests on the recollection of whoever was in the room.

Bottom line: an interview scorecard is not a form, it is an agreement about what is being measured, and its value is set almost entirely by when you write it. Written the week of the onsite, it makes a late-stage conversation tidier. Written the day the requisition opens, it decides who enters the funnel, what each interview is for, and whether the numbers coming out the other end mean anything. Same document, completely different instrument. Score your next applicant batch against your role criteria with Rankid and let the scorecard start where the candidates do.