Structured Interviews and Reducing Bias

Scenario

Example organisation: interviewers at the 40-engineer company "go with the flow" — each conversation is different, follows the candidate's CV, and ends with a gut feeling. Two candidates with similar experience get opposite results depending on who interviewed them. The team notices it hires people who resemble the existing team.

What you see on the team

• Inconsistent questions and outcomes across interviewers.
• Feedback like "great culture fit" or "didn't feel senior".
• Debriefs dominated by first impressions.
• A team that becomes less diverse with every hire.

The common mistake

Running unstructured interviews — different questions for each candidate, judged by gut feeling — and assuming experienced interviewers can "just tell".

Why it happens

Unstructured interviews feel natural and personal; interviewers trust their instincts. But instincts reward familiarity and confidence, not job performance. Google's re:Work guide on structured interviewing notes that structured interviews result in increased predictive validity and decreased differences between demographic groups, and that structured assessment tools ("rubrics") are more predictive than unstructured interviews.

The playbook

Plan questions in advance for each competency on the scorecard. Use two kinds, as the re:Work guide describes:

• behavioral — "Tell me about a time you led a migration across teams. What was hard?"
• hypothetical — "A product team insists on a deploy freeze exception during an incident. What do you do?"

Write a rubric per question with examples of poor, mixed, good and excellent answers:

Competency: leading projects across teams
Q: Tell me about a time you led work that depended on other teams.
 Poor:      describes own work only; blames other teams for delays
 Mixed:     coordinated, but dependencies discovered late; no plan B
 Good:      mapped dependencies early, agreed dates with owners, tracked visibly
 Excellent: also reduced the dependency (self-service, contract-first) and
            communicated risks to stakeholders with options

Same questions, same order, same rubric for every candidate for the role.

Take notes on evidence, score independently before the debrief, then discuss.

Train interviewers, including shadowing and reverse-shadowing before they interview alone.

Refresh questions periodically — the re:Work guide notes candidates compare notes.

What to say

Interviewer briefing:

"Use the three planned questions for your competency, in order. Write down what the candidate said, not your impression. Score against the rubric before you read anyone else's feedback."

Trade-offs

• Structured interviews can feel stiff; leave a few minutes for the candidate's questions and natural conversation.
• Writing rubrics takes time; reuse them across hires for the same role.
• Rubrics don't remove all bias; diverse panels and independent scoring still matter.

Signals it's working

• Interviewers' scores agree more often in debriefs.
• Feedback cites evidence ("described mapping dependencies at planning"), not impressions.
• The candidate pool and hires become more diverse.
• New hires perform as the interviews predicted.

Key points

• Structured interviews predict performance better and reduce demographic differences in outcomes.
• Plan behavioral and hypothetical questions per competency; write a rubric with example answers.
• Same questions and rubric for every candidate; score independently before the debrief.
• Train interviewers and refresh questions over time.