Interview Process and Scorecard Template

List what the new hire must be able to do. Start from the job description's must-haves or, better, a requirement register that says how each one would be checked. The US Office of Personnel Management (OPM) notes that interview questions built on competencies a job analysis identifies as critical to success¹ show high validity and strong agreement between raters.

Next, sort every requirement into one of three groups.

• Test it in the interviews: skills you can only judge by watching or asking, such as training a room of new staff.
• Confirm it from a document: facts such as a license, a degree or a stated willingness to travel. Check these once at screening and give them no interview time.
• Rewrite it or drop it: anything nobody can describe in observable terms. "Culture fit" can't be rated until you name the behavior you mean.

Give each skill a weight so the weights add up to 100. Base each one on how much of the job the skill covers and what a weak hire on it would cost. The must-haves together should outweigh the nice-to-haves. Then date the sheet.

Weights matter because people gather evidence well and combine it badly. Writing in Harvard Business Review, researchers Nathan Kuncel, Deniz Ones and David Klieger call people very bad at weighing the results² of a hiring process. In the 17 studies they analyzed, a simple equation beat human judgment by at least 25%.

Set before the first interview, the weights give your panel a simple equation of its own. Leave them alone once anyone has been rated: a weight changed after the panel has a favorite is a justification, not a weight.

Structured questions, the same for every candidate, test most skills. Ask about the past ("Tell me about a project you ran that slipped") or pose a situation from the job ("A customer's launch will be two weeks late. What do you tell them?"). A behavioral interview question bank gives you questions and rating anchors grouped by skill.

When a skill is needed from day one, use a work sample, such as reconciling two real files with one planted error. OPM describes work samples as tasks that mirror what employees do on the job³. It also warns they can be costly to develop and may not fit skills you plan to teach after hiring.

Both methods earn their place. Writing for SIOP, the professional society for workplace psychologists, two researchers summarize a 2022 reanalysis of hiring research by Paul Sackett and colleagues. The reanalysis found that structured interviews predicted job performance best, with work sample tests also in the top five⁴.

Put one interviewer's name against each skill. That person prepares the test, runs it and writes the rating. Without that, interviewers ask about whatever needs no preparation. Easy topics often come up three or four times, and skills that need a prepared test get none.

Allow one deliberate overlap: a blind second rating on the heaviest skill, made without seeing the first and written into the plan with its reason. Anything else an interviewer wonders about goes in their debrief notes.

Budget the time next. Divide each skill's minutes by its weight. A 5-point skill booked for an hour, or a 25-point skill with no minutes, means the schedule and the weights disagree.

The scorecard needs one row per skill: its weight, its owner, the rating, the level the answer reached and the evidence. Write the levels before the first interview. Google's re:Work site says the company often documents, with examples, what a poor, borderline, solid and outstanding answer would cover⁵ for each skill a question tests.

Four levels with no middle option make each interviewer come down on one side. Every rating also needs its evidence: what the candidate said or did, close to their own words. Treat a rating without evidence as missing, not low: it leaves that skill untested.

Ask every candidate the same questions in the same order. Iris Bohnet, a behavioral economist at Harvard Kennedy School, advises scoring each answer right after it's given⁶, since we're more likely to remember vivid and recent answers. She also suggests comparing all candidates' answers to one question before moving on to the next.

Interviewers rate alone and submit their scorecards before the debrief. In the meeting, read the weighted totals once, then spend the time on disagreements and on ratings with thin evidence. A hiring debrief tool can pull the real disagreements and unsupported verdicts out of a panel's scorecards.

Keep the plan and every scorecard. In the US, an EEOC rule requires covered employers to keep hiring records for one year⁷ from when they are made or from the decision, whichever is later. A discrimination charge extends that until the case is closed.

The template on this page lays out that whole plan, from the assignment sheet to the debrief format. Open it in River, and River starts by measuring what your current interviews already cover.