Founders Feed

Use an Interview Evidence Card Before the First Candidate Arrives

Define the outcomes, evidence, scoring anchors, interview owners, and decision record before candidate chemistry starts moving the standard.

Saturday, September 26, 2026

4 min read

Operator Intelligence, Hiring, Frameworks

The Signal

A candidate leaves the call and the team says they were sharp, confident, strategic, or a good fit. Those reactions may be real. They are not yet hiring evidence.

The evidence is what the candidate did, the conditions they worked under, the decision they made, the result they produced, and what they learned when the result missed. If the team cannot connect its recommendation to the work the role must own, the interview has become a memory contest.

An interview evidence card fixes the standard before chemistry starts moving it. The card names the role outcomes, questions, scoring anchors, interviewer owner, and evidence required for a decision. It gives every interviewer the same operating object to complete.

Why this matters now

Hiring conversations are easier to record, summarize, and circulate. That can save time. It can also make a weak process look stronger because the notes are cleaner. A detailed summary does not help if each interviewer asked different questions or used a private definition of strong.

The risk gets larger in founder-led companies. The role may be new. The founder may be carrying part of it today. One interviewer values speed, another values polish, and a third is screening for experience with a tool that may not matter six months from now. The candidate is compared against several jobs without anyone admitting the job changed.

The U.S. Equal Employment Opportunity Commission says employment decisions should not be based on protected characteristics and warns against neutral practices that disproportionately exclude protected groups when they are not job-related and necessary. This is not a substitute for legal review. It is an operating reason to define job-related criteria and apply them consistently before the team starts choosing favorites.

The mistake to avoid

The mistake is adding a scorecard after the interviews. Once the team likes a candidate, vague criteria become easy to bend. The scorecard turns into documentation for a decision already made.

Do not solve that by turning the interview into a rigid script with no follow-up. The fixed part is the outcome and scoring standard. The interviewer should still ask for detail, test contradictions, and understand context. Human judgment belongs in the process. It should not be allowed to silently change the job.

Another mistake is scoring personality words. “Executive presence,” “culture fit,” and “high energy” are too elastic unless the role requires a specific observable behavior. Replace the label with the work: can this person explain a missed forecast, resolve a customer escalation, build a weekly plan, or make a decision with incomplete data?

Run the interview evidence test

Take the last three candidates considered for the same role. Put their interview notes side by side. Mark every question that was asked of all three, every answer tied to a stated role outcome, and every score supported by a concrete example.

Then inspect the gaps. Did one candidate receive a useful follow-up that another did not? Did an interviewer introduce a new requirement halfway through the process? Did the group discuss confidence more than results? Did anyone score a candidate before the interview notes were complete?

Use three anchors for each outcome. Weak evidence means the candidate cannot show ownership or a relevant result. Acceptable evidence means they handled a comparable problem with a clear role and result. Strong evidence means they can explain the trade-off, operating method, result, failure, and adjustment without borrowing credit from the team around them.

What stays protected

Protect the actual role. Do not let a familiar company name, polished answer, or tool certification replace the outcomes the person must own. Keep compensation, reporting line, decision rights, travel, schedule, and location requirements consistent with the approved role.

Protect candidate information. Interview notes should stay in the authorized hiring system with appropriate access and retention, not in personal notebooks or open chat threads. Do not feed recordings or candidate data into an AI tool unless the company has approved the tool, disclosure, access, and retention path.

Protect independent judgment until the evidence is recorded. If the most senior person announces a verdict first, the rest of the panel will often score toward it. Have each interviewer submit evidence before the group review.

The first move

Choose one open role. Write five outcomes the person must own in the first year. For each outcome, create one primary question, two allowed follow-ups, and weak, acceptable, and strong evidence anchors.

Assign one interviewer to each outcome. Remove duplicate questions. If an outcome has no owner or cannot be scored from observable evidence, the role is not ready for another interview yet.

The move this week

Run the card with the next candidate and require every interviewer to submit evidence before the debrief. In the meeting, discuss one outcome at a time. Separate missing evidence from negative evidence. A question the team failed to ask is not a candidate failure.

After the decision, keep the completed card with the authorized hiring record and review where the process broke. Tighten one question, scoring anchor, or owner before the next interview.

The objective is not to remove judgment from hiring. It is to make sure judgment is about the work, uses the same standard, and leaves a record the company can defend and improve.

Brian Stewart at his desk on a video call, explaining with both hands
September 1, 2026

Learn it live with me

Personal AI Agent

I set up a personal AI agent live, start to finish, on standard off‑the‑shelf tools.

When
Tuesday, October 20, 2026, 6–8 pm CT
Length
About 2 hours, on Riverside
Price
$47, one time
Replay
Included, by email within 48 hours, and kept on brianwith.ai for 12 months.