Let's talk about how hiring actually happens in most Nigerian companies. A CV arrives through somebody's cousin. The MD "chats with" the candidate for twenty minutes, asks "can you work under pressure?", likes their spirit, and makes an offer. Three months later everyone is wondering how such an impressive interviewee turned out to be such a disappointing employee.
Here's the uncomfortable truth: the standard Nigerian interview — an unstructured chat where each panellist freestyles whatever questions enter their head — is barely better than tossing a coin. And in this market, where a wrong hire means months of salary, NYSC-style retraining from scratch, and another round of recruitment, you cannot afford coin-toss hiring.
The good news is that interviewing well is not a talent. It is a process. And the process is boringly learnable.
Why "I'll know them when I see them" keeps failing you
Decades of selection research, including the famous Schmidt and Hunter meta-analysis of 85 years of hiring studies, found that structured interviews are among the strongest predictors of actual job performance — clearly ahead of the unstructured "let's just gist" interview. More recent re-analyses published in the 2020s went further: when researchers corrected the statistics, structured interviews came out as the single best-performing selection method commonly available to employers.
What's a structured interview? Simple: every candidate for the role gets the same questions, in the same order, scored against the same written standard. That's it. No magic.
Unstructured interviews fail for a predictable reason: interviewers decide in the first few minutes based on confidence, accent, appearance, or shared connections — then spend the rest of the interview collecting evidence for a decision they've already made. The candidate who schooled abroad and speaks with polish wins the chat. The quietly excellent one who has actually done the job loses. Your company pays for that bias every month.
Start before the interview: define what "good" looks like
Most interview failures happen before anyone enters the room, because nobody agreed what the job actually requires. Before you shortlist a single CV, write a one-page scorecard for the role:
The 3 to 5 outcomes this person must deliver in their first year. Not activities — outcomes. "Reduce customer response time from 24 hours to 4 hours," not "handle customer service."
The 4 to 6 competencies that drive those outcomes. For an accountant at a Lagos trading company that might be: reconciliation accuracy, tax compliance knowledge, Excel skill, integrity under pressure, and communication with non-finance people.
The genuine deal-breakers. Be honest and few: if the role truly requires resumption at 7am on the Island, say so now, not after the offer.
Every interview question you ask should trace back to this scorecard. If a question doesn't test a competency on the sheet, it's conversation, not assessment.
Write your questions down. Yes, all of them.
For each competency, prepare two kinds of questions.
Behavioural questions ask about the past. "Tell me about a time a major client was furious and it was genuinely your company's fault. What did you do?" Past behaviour is the best evidence of future behaviour, and candidates find it much harder to fabricate a detailed, coherent story under follow-up questions than to recite prepared lines.
Situational questions pose a realistic scenario. "It's month-end close. NEPA has taken light, the gen is down, and your CFO needs the reconciliation by 8am. Walk me through your next three hours." You're listening for how they think, prioritise and communicate — not for a heroic speech.
Probe every answer with follow-ups: What exactly was your role? What did you say? What happened next? What would you do differently? Vague answers dissolve under three follow-ups; real experience gets richer.
And retire the tired ones. "Sell me this pen" tells you nothing about a data analyst. "What's your greatest weakness?" produces rehearsed theatre. Questions about marital status, pregnancy plans, tribe or religion are not just useless for predicting performance — they expose you to discrimination complaints and quietly tell your best female candidates to take their talent elsewhere.
Score like an examiner, not a fan
After each interview — before any hallway gist — every interviewer independently scores each competency from 1 to 5 against written anchors. A "5" on client handling might read: "gave a specific example with a measurable outcome, owned their mistake, described what they changed afterwards." A "2": "spoke in generalities, blamed others, no concrete example."
Two rules make this work. First, score immediately — memory contaminates fast, and by candidate four everyone blurs together. Second, score independently before discussing, because once the oga at the top says "I liked her," every junior panellist's score mysteriously rises. If two interviewers disagree wildly on a competency, that's not a problem — that's the process surfacing something worth discussing with evidence instead of vibes.
