Two strong finalists means the process worked. That’s the good news. The bad news is that the final decision usually comes down to gut feel, which is exactly what the structured process was designed to replace. The comparison doesn’t end with the interviews; it ends with the evidence.

The interview kit gave you consistent answers. The reference check surfaced what the interviews didn’t. The scorecard has each candidate’s performance against each criterion. Read it rather than setting it aside for a feeling.

This article is the last step in the process built across The Reference Check System That Surfaces What the Interview Didn’t, the interview kit, and the scorecard. It is how to use all of it to make a decision that is both confident and defensible.

The Evidence-Based Hiring Decision has three steps. None of them are complex. Together they produce a decision that the evidence supports.

Why the Final Decision Defaults to Gut Feel

The obvious reason: two strong candidates look similar on paper and similar in the interviews. Neither has a clear flaw. The decision is genuinely close. Gut feel is the default when evidence doesn’t produce a clear winner.

The less visible cost is what gut feel is actually measuring at this moment. The final impression is dominated by recency. Whoever interviewed last is most vivid, and whoever said something memorable carries more weight than the scorecard justifies. The candidate who was slightly nervous in one question loses ground the scorecard doesn’t support.

The deepest problem is that gut feel at this stage is invisible to the business. When the hire doesn’t work out, you cannot identify what you were measuring when you made the decision. There is no record of the reasoning, no way to calibrate what to look for next time, and no pattern to read across decisions.

A final decision made from the evidence record can be explained. A final decision made from gut feel usually can’t be.

The Evidence-Based Hiring Decision

The three steps use the evaluation record built across the search to produce a comparison that doesn’t rely on impression.

Step 1 reads the scorecard before the decision conversation

Before any discussion about the two finalists, each decision-maker reads the scored criteria independently. Not a discussion, not a meeting. Each person reads the scorecard for each candidate and notes where the two candidates differ most.

This step prevents the dominant voice in the room from anchoring the comparison. When everyone arrives having read the evidence independently, the discussion is about what the evidence shows rather than what the most confident person feels.

Before: The decision conversation starts from “what did you think?” and goes from there. After: The decision conversation starts from “here’s where the scorecard shows they differ, let’s talk about those gaps.”

Step 2 compares the gaps, not the totals

Two candidates with similar overall scores may differ significantly on the criteria that matter most for this specific role. Step 2 focuses the comparison on those gaps rather than trying to produce a total score.

Which criterion is the single most important for success in the first six months? Find the gap between the two candidates on that criterion first. If one candidate leads on the most important criterion and the other leads on less critical ones, the decision isn’t as close as it appeared.

If the gap on the most important criterion is genuinely small, look at where each candidate’s specific concern shows up in the record. Which concern is more manageable given what this role actually requires? Which candidate’s weakness would produce less friction in the first ninety days?

Step 3 reads the reference check finding against the scorecard

The final step is the comparison between what the interview showed and what the reference check confirmed or contradicted.

If the reference check aligned with the scorecard on the top candidate, the decision is confirmed from two sources. If the reference check revealed something different from what the interview suggested, that gap is the most important thing to weigh before deciding.

A candidate whose references confirm the scorecard is a different proposition than one whose references introduced a concern the interview didn’t surface. The reference check is not a veto. It is additional data that makes the comparison richer.

How the Conductor Surfaces Your Final Decision

When the decision between two strong candidates is genuinely close, the Conductor can read the accumulated record and return a comparison the gut can’t produce.

The Living Library holds the scorecard, the interview notes, and the reference check summary for each candidate. The Conductor, Kiluma’s context-aware AI, reads from that record rather than from a generic comparison framework. Ask it: “Based on the scorecard and reference records, where are the most meaningful differences between these two candidates?” It reads the criteria, ratings, concern notes, and reference summaries and returns the comparison from the evidence, not from impression.

That’s not a recommendation to hand the decision to the Conductor. The decision is yours. The Conductor returns the evidence organized around the role’s criteria so the decision comes from what the process produced, not from the last interview’s impression.

Read the Scorecard Before the Decision Meeting

Before the next decision conversation, set a rule: everyone reads the scorecard for both candidates before the meeting starts. No discussion until everyone has read it.

That single step changes the conversation from “what did you think?” to “what does the record show?” The gap between those two conversations is the gap between a structured decision and a gut-feel one.

A Different Way to Hire

Before the work of this chapter, the hiring decision for two strong finalists ended with the person who interviewed last or the candidate who told the best story in the final round. After it, the decision ends with the evidence:

The before and after are not the same process. The before hired whoever felt strongest; the after hires whoever the evidence supports. Only one can be explained. Try Kiluma free for 14 days at kiluma.ai.