
Two finalists remain. The CEO favors A’s campaign experience; the marketing lead prefers B’s work sample analysis. After an hour of circular debate, the CEO says, “Let’s go with A.” Three months later, someone asks why. There’s no record to explain the decision.
When opinions split, start with the evidence for each agreed hiring criterion. Put findings from resume screening, work samples, and interviews side by side. Separate what you observed from what you inferred and what you haven’t verified. That gives you a basis for deciding whether to make an offer or ask another question.
Consider a skincare brand with roughly 50 employees that sells through its own online store and is hiring an experienced performance marketer. Finalist A described campaign experience in detail during an experience-based interview. Finalist B compared two channels using first-time customer data in an advertising performance work sample. Their hiring debrief shows how to fill out an evidence comparison table and decision record. Both candidates are assessed against three behaviors from the hiring scorecard, using records from the work sample task and experience-based interview.

Open with “Who do you like better?” and the first speaker’s impression, or the most senior person’s opinion, can become the yardstick for everyone else. Before the meeting, review the work this person will own and the behaviors you expect. Check that each interviewer has completed an independent evaluation.
The U.S. Office of Personnel Management (OPM) recommends this sequence in Structured Interviews: A Practical Guide (2008). Panel members should review their notes and score candidates before discussing them with colleagues. They then compare notes, scores, and supporting observations, examine why their ratings differ, and reach agreement. For each score, the guide recommends recording what the candidate actually said and why that response warrants the rating.
Here, the three criteria are distinguishing verified numbers from assumptions, assessing how budgets and discounts affect new customers and profit, and checking before launch. Keep a candidate’s previous advertising budget, employer’s brand name, and speaking ability separate from behavioral evidence against those criteria. If a required criterion remains unverified, mark what’s missing instead of assigning a low score. If you change a criterion during the debrief, document why and what you’ll need to verify with candidates who have already completed the process.
An evidence table starts with each interviewer’s own records. Google’s re:Work, which shares practical HR resources, also recommends taking detailed notes on candidate responses and revisiting them during hiring in its structured interviewing guide. The table below organizes A’s and B’s records under the same criteria in matching columns.
| Shared Criterion | Candidate A | Candidate B |
|---|---|---|
| Distinguishes Facts From Assumptions | Task: Treated ad-attributed revenue and ROAS as new customer results Interview: Whether repeat purchases were separated is not yet verified | Task: Did not rely on ROAS that included repeat purchases Interview: Described rechecking performance metrics |
| Assesses New Customer and Profit Impact | Task: No comparison based on first-time customers Interview: Described setting campaign goals and defining target customers | Task: Compared ad spend per customer; did not calculate profit after discounts Interview: Experience assessing profit was not covered |
| Checks Before Launch | Task: Committed to a 15% coupon without checking with the operations lead Interview: Described coordinating a promotion with the operations lead in detail | Task: Proposed checking with the operations lead before offering a discount; did not address budget approval limits Interview: Own contribution to a campaign run with the marketing lead was unclear |
Your next step depends on whether interviewers interpreted the same answer differently, observed different situations, or lack information altogether. For conflicting interpretations, compare the candidate’s exact response with the criterion. For conflicting evidence, ask about the context and reasoning behind each choice. For missing information, first decide whether further verification is needed.
| What Comes Up in the Debrief | Type of Difference | Next Step |
|---|---|---|
| “Writing that they would check with the operations lead is enough.” / “It’s unclear what they personally did.” | Different interpretations of the criterion and individual contribution | Compare the exact response with the expected behavior. |
| “They described coordinating a promotion in detail during the interview, but committed to adding a coupon without checking in the task.” | Conflicting evidence across assessments | Use the same criterion to investigate why their choices differed. |
| “We didn’t ask about their experience assessing profit after discount costs.” | Information not yet verified | Depending on its importance, record a follow-up question or a condition for the decision. |
Missing information can’t justify an endless series of interviews. First ask whether the answer could change your decision and whether it already exists in the record. Define the behavior you need to verify and how you’ll judge it before writing the question. That also reduces the burden on candidates to repeat themselves.

For A: “Why did you allocate the entire budget to channel A based only on ROAS? How would you change the allocation if ad-attributed revenue included repeat purchases by existing members?” For B: “In the campaign you ran with your marketing lead, what did you personally propose and execute?” The wording differs, but both questions serve the same purpose: filling gaps against the same hiring criteria.
| Debrief Step | What to Record |
|---|---|
| Review the hiring purpose and criteria | The role and behaviors needed from day one |
| Share independent evaluations | Evidence for each candidate and reasons for ratings |
| Examine differences | Differences in interpretation, conflicting evidence, and unverified items |
| Decide or verify further | Decision-maker, rationale, open questions, and next owner |
| Hand off after hiring | Initial assignments, support, and observation plan |
Suppose B gives a detailed follow-up answer explaining what they personally initiated and executed in the campaign run with their marketing lead. Add that response to the record as new evidence before making a judgment. If an important criterion remains unverified, document the gap and decide whether to investigate further or put the decision on hold. If you choose to offer B the job, record the supporting evidence: B compared ad spend per first-time customer and proposed checking with the operations lead before offering a discount. Keep profit after discount costs and budget approval limits marked as not yet verified. For B’s first promotion after joining, plan to observe whether they account for discount costs in profit calculations and check budget approval limits upfront. The OPM guide explains that interview notes should document the reasoning behind each score and support the hiring decision. Recording what you knew and didn’t know at the time helps prevent you from judging the new hire by a different standard later.

You can add interview response evidence organized by Interview Pro and evidence from choices and answers in Assessment to the relevant rows of the comparison table. Identify the underlying source for each entry. Don’t average scores that use different scales.
When reviewing Telta’s Interview Pro overview and Assessment overview, look beyond the total score to see which criterion each piece of evidence supports. After the hire, use the quality-of-hire review checklist to connect expected assignments with observations of actual work. Also decide what support and opportunities onboarding will provide to assess behaviors you couldn’t verify during hiring.
Start with what each method assessed and under what conditions. Compare the candidate’s actual responses and work against the shared criteria, alongside anything still unverified. If the difference matters to the decision, investigate why it exists.
Not observing a behavior is different from observing poor performance. Mark unverified items separately so the team can decide whether the criterion is required and whether further assessment is needed.
Related Reading