Two candidates meet the same panel. One receives careful follow-up questions. The other is interrupted, rushed and asked about a different set of topics. At the end, the panel calls the process fair because both people sat in the same room.

That is not consistency. It is two different assessments carrying the same job title.

A structured interview gives every candidate a comparable opportunity to produce relevant evidence. It still feels human. The difference is that warmth no longer decides which evidence the panel remembers to collect.

Define the evidence before writing questions

Start with four to six capabilities that are genuinely important in the role. For each one, describe what good performance looks like in observable terms.

For example, “handles customer conflict” could mean:

  • Listens before proposing a solution.
  • Separates the immediate complaint from its underlying cause.
  • Explains the next step and records the issue.
  • Escalates when the decision falls outside their authority.

These behaviours become scoring anchors. They are stronger than labels such as confident, polished or “a good fit,” which can mean something different to every interviewer.

The Keele University training video models a panel using a person specification and asking about organisation, busy teamwork, communication and complaint handling. Its value is not a set of magic questions. It shows how role requirements can guide the conversation.

Ask the same core questions

Give every candidate the same core question for each capability, in the same order where practical. A behavioural question might be: “Tell us about a time you had to resolve an unhappy customer’s problem while working within a policy.”

A situational question might be: “A customer says your team promised a delivery today, but the order is still at the warehouse. What would you do first?”

Behavioural questions explore past evidence. Situational questions reveal how someone reasons through a defined scenario. A balanced interview can use both, provided each question connects to the scorecard.

Do not add a harder question because one CV looks unusual or an easier one because a candidate was referred by a colleague.

Probe without supplying the answer

Structure does not mean reading a script without listening. Panels can ask standard probes when an answer lacks detail:

  • What was your responsibility?
  • What did you do personally?
  • What information did you use?
  • What happened next?
  • What would you change now?

Avoid a leading probe such as, “So you spoke to the manager and documented it, right?” That gives the candidate the behaviour you wanted to hear.

Also allow a short pause. Some people answer quickly because they have rehearsed. Others need a moment to retrieve a genuine example. Thinking time is not evidence of weak ability.

Take notes about actions, not personality

Interview notes should capture what the candidate said and did in the example. “Checked the invoice, called the warehouse, gave the customer a revised time and logged the error” is useful. “Excellent communicator” is a conclusion without a trail.

Tell candidates that the panel will take notes so the silence does not feel strange. Divide responsibilities: one person can chair, another watch time and all panel members can record evidence against assigned criteria.

Keep notes secure and collect only information needed for the hiring purpose. Application and interview records contain personal information, so a South African employer should align access, retention and disposal with its privacy responsibilities.

Score before anybody persuades the room

Use a short anchored scale. A middle score might mean the candidate supplied a relevant example but needed several probes or missed one important action. A high score should require clear, specific evidence across the agreed behaviours.

Immediately after the interview, each panel member should score independently and write a reason. Discussion starts after those scores exist.

When scores differ, ask what evidence produced the difference. Do not average away a serious concern or let seniority settle the debate. Return to the anchor and the candidate’s answer.

The final record should explain why evidence met or did not meet the requirement. It should not contain comments about accent, appearance, family circumstances or other characteristics unrelated to performing the job.

Calibrate before and after the round

Before the first candidate, give the panel a sample answer and ask everyone to score it. Differences reveal vague anchors while there is still time to fix them.

After the round, review whether a question produced useful evidence. If candidates repeatedly misunderstood it, the problem may be the question. If one interviewer consistently scored more harshly, compare their interpretation with the agreed anchors.

Do not rewrite the scoring rules after meeting a preferred candidate. Improvements belong in the next round unless the current process is clearly invalid or unfair, in which case pause and correct it consistently for everyone still under consideration.

Keep the experience recognisably human

Explain the format at the start. Introduce the panel, confirm the available time, offer reasonable accommodations and leave space for the candidate’s questions. A structured process should reduce uncertainty, not create a courtroom.

Close by explaining the next step and realistic timing. Then keep that promise.

The goal is not to turn people into numbers. It is to stop an unrecorded feeling from becoming the number that decides their future. A thoughtful structure helps interviewers listen more carefully because they know which evidence they need and why it matters.

Sources and limits

The video is a training demonstration, not evidence that its exact panel or questions fit every role. South African employers should obtain appropriate HR or legal advice for their circumstances. This article is an original synthesis, not a transcript.