You have sat in this debrief. Four people interviewed the same candidate. Two of them are certain. Two of them are unsure and cannot say why. Somebody says "strong culture fit," everybody nods, and a hire that will cost a year of salary gets decided in eleven minutes on the strength of a phrase nobody defined.
The fix is not more interviews. It is a scorecard: a short, written definition of what you are measuring, agreed before anyone speaks to a candidate, and filled in independently by every interviewer afterwards.
Confidence and accuracy are different things. Interviewers who improvise usually come away feeling they read the person well — and the research says they read them worse. Meta-analytic work on employment interviews has consistently found that structured interviews predict job performance substantially better than unstructured ones, with recent estimates putting structured formats at roughly double the predictive validity of a free-form conversation (Wingate, International Journal of Selection and Assessment, 2025; see also the long-standing McDaniel et al. meta-analysis, 1994).
Structure also narrows the gap where bias lives. A meta-analytic comparison of the two formats found both were affected by bias, but that unstructured interviews were markedly more susceptible to it ("Do structured interviews eliminate bias?"). Structure does not eliminate bias. It reduces the surface area.
A working scorecard fits on one screen. It has four parts:
Four to six criteria, no more. Each one is a capability the job genuinely requires — not a personality trait. "Can reconcile a month-end ledger under deadline" is a criterion. "Detail-oriented" is a horoscope.
A rating scale with words attached. A 1–4 scale where each number has a written description of what that level of evidence looks like. Four points, not five — an odd-numbered scale invites everyone to park in the middle.
An evidence box per criterion. The interviewer writes what the candidate actually said or did. A rating without evidence is an opinion wearing a number.
One overall recommendation. Hire, no hire, or hire for a different role. Not a percentage.
Every criterion needs its own question, asked the same way to every candidate, with follow-ups planned in advance. Past-behaviour questions ("walk me through the last time you…") outperform hypotheticals, because hypotheticals reward people who are good at interviews rather than people who are good at the job.
This matters more in our region than most hiring playbooks admit. A finance lead in Nairobi may have run a business unit that never had a formal title attached to it. A logistics coordinator in Dubai may have three years of Gulf experience under a job title that means something different in every company that used it. A great candidate may have spent two years in the informal sector doing work no CV format has a box for. If your criteria are written as capabilities and your questions ask for evidence, all of that becomes visible. If you are scanning for brand-name employers and tidy title progressions, you will filter out your best people and never know it.
The same principle applies to language. A candidate interviewing in their second or third language will often sound less fluent than they are competent. Score the evidence, not the polish.
Most teams do the work and then throw it away at the last step: they open the debrief by asking the most senior person in the room what they thought. Everyone else's independent judgement quietly reshapes itself around that answer.
Instead: every interviewer submits their scorecard before the debrief starts, and nobody sees anyone else's until all are in. Then read the ratings out criterion by criterion and spend the meeting on the disagreements — those are the only part worth discussing. Where two interviewers split on the same criterion, the question is almost always "what evidence did you each hear?", and answering it usually resolves the split in under two minutes.
Before your next opening goes live:
1. Write four to six job-specific criteria with the hiring manager, before the job ad is published.
2. Define each point on a 1–4 scale in plain words.
3. Assign each criterion to a specific interviewer so nobody duplicates and nothing is missed.
4. Write the primary question and two follow-ups for each criterion, and use them verbatim with every candidate.
5. Require written evidence for every rating.
6. Collect all scorecards before the debrief opens, and start with the disagreements.
None of this requires software. It requires the discipline to decide what "good" means before you meet someone who is charming.
Where a platform helps is in making the discipline the default rather than an act of willpower: scorecards attached to the role inside your applicant tracking system, structured assessments that generate comparable evidence before the interview stage, and video interviews where every candidate answers the same questions and the panel scores independently. That is the shape TalentBridge AI is built around — consistent evidence in, human judgement on top.
See how TalentBridge AI can help your business hire better, faster, and more efficiently.