Purpose

Unstructured hiring is expensive. It produces inconsistent outcomes, introduces significant bias, and creates decisions that cannot be defended after the fact. When a hire does not work out and the manager is asked "what did the interview process show you?" they should be able to answer specifically.

A hiring loop solves this by:

• Defining upfront what you are assessing and why
• Distributing that assessment across interviewers so no single person has to assess everything
• Using consistent, evidence-based scoring so debrief comparisons are meaningful
• Removing redundancy so candidates are not asked the same questions five times
• Producing a decision that is documented, defensible, and not just based on who argued loudest in the debrief

Done well, a hiring loop:

• Assesses the criteria that actually predict success in the role
• Gives every candidate a fair and consistent experience
• Produces a clear, documented outcome - hire or no hire, and why
• Moves fast enough that strong candidates do not accept other offers while you deliberate

Done badly, it is a series of conversations where interviewers ask whatever they feel like, nobody compares notes until the end, and the decision is made by whoever has the strongest opinion in the debrief.

When to Use This Playbook

Use this playbook when:

• You are opening a role and need to design the interview process
• You have an existing process that is producing inconsistent or unexplainable outcomes
• You are adding new interviewers and need to bring them up to standard
• You want to improve the speed of your hiring process without reducing quality
• You have recently had a bad hire and need to understand what the process missed

This is not the right tool for:

• Sourcing and attracting candidates (that is a separate exercise)
• Designing the job description (that comes before this)
• Onboarding the person once hired (see the onboarding playbook)

Before You Start

Before you design the loop, you need two things: a clear role profile and a scorecard.

The role profile

The role profile answers: what does success in this role look like at 6 months, 12 months, and 24 months? Not a list of skills. A picture of the impact.

Write it in outcome language:

• "At 6 months, this person has shipped the data pipeline refactor and is unblocking the data science team."
• "At 12 months, they are the go-to for distributed systems architecture decisions in the squad."
• "At 24 months, they are mentoring two mid-level engineers and running the technical strategy for the platform."

The role profile is used by interviewers to set context and by the debrief to evaluate fit.

The scorecard

The scorecard defines the criteria you are assessing and what evidence looks like at each level of that criterion. Every interviewer uses the same scorecard. Every score must be backed by evidence.

A good scorecard for an engineering role has four to six criteria. More than six creates noise. Fewer than four is probably missing something important.

Example criteria for a senior engineer:

Each criterion is scored 1-4:

A "strong hire" typically needs a 3 or above on all criteria, with no 1s. A 4 on most criteria with a 2 on one is usually still a strong hire if the 2 is in an area the role does not heavily depend on.

The Process

Step 1 - Design the loop

The loop has four to five stages. Every stage has a defined purpose and a named interviewer. No redundancy - if two stages are assessing the same thing, remove one.

A standard loop for a senior engineer:

Assign each interviewer to specific criteria from the scorecard before the loop begins. The technical interviews cover technical depth and problem decomposition. The behavioural interview covers collaboration, delivery, and growth. The hiring manager covers motivation and leadership.

Step 2 - Brief every interviewer before the loop starts

Every interviewer should know:

• What criteria they are responsible for assessing
• What the role is and what success looks like
• What questions they are not allowed to ask (protected characteristics - see bias section)
• How to score using the scorecard
• When they need to submit their scorecard (ideally within two hours of the interview)

The briefing is 20 minutes per interviewer. It is not optional. An interviewer who has not been briefed should not be in the loop.

Template for the interviewer briefing:

"In this interview, your job is to assess [criteria A] and [criteria B]. You are not responsible for the technical assessment - [other interviewer] is covering that. Your role is to understand whether [candidate] can [specific thing]. Here is what strong evidence looks like for this level. Here is the scorecard you will complete. Please submit it within two hours of the interview ending."

Step 3 - Run the interviews

The STAR format. All behavioural interviews should use the STAR format for follow-up questions. When a candidate gives a vague answer, probe it.

STAR = Situation, Task, Action, Result

If a candidate says: "I've led cross-functional teams and improved delivery speed."

That is not STAR. Ask:

• "Can you give me a specific example of a time you did that?"
• "What was the situation - what was the team, the goal, the context?"
• "What did you specifically do - not the team, you?"
• "What happened as a result? How did you measure it?"

The gap between what candidates say they did and what they actually did is usually large. STAR closes it.

Structured questions. Every interviewer for a given criterion should ask the same core questions. This makes debrief comparisons meaningful. If one interviewer asked about conflict and another asked about mentoring, you cannot compare their assessments.

Core behavioural questions by criterion:

Delivery and ownership:

• "Tell me about a project that nearly failed and what you did about it."
• "Give me an example of something you shipped that you are proud of. What was your specific contribution?"
• "Tell me about a time when you committed to something and then could not deliver it. What happened and what did you do?"

Collaboration and communication:

• "Describe a time when you disagreed with a technical decision that was made. What did you do?"
• "Give me an example of explaining a complex technical concept to a non-technical stakeholder. What did you do and how did it land?"
• "Tell me about a time a team member was struggling. What did you do?"

Growth and learning:

• "What is the most significant technical thing you have learned in the last 12 months? How did you learn it?"
• "Tell me about a time you got feedback that was difficult to hear. What did you do with it?"
• "What do you currently find most challenging about the level above yours?"

Technical depth (in a design interview):

• "Walk me through how you would design [system X] - start with requirements."
• "What trade-offs did you consider and why did you choose this approach?"
• "What would you change if the scale requirement was 100x larger?"
• "What are the failure modes of the design you've described?"

What to do when the interview is going badly.

Sometimes it is clear within 20 minutes that the candidate is significantly below the bar. Do not fake enthusiasm for the remaining 40 minutes. You can:

• Redirect to confirm the assessment: try a different angle on the criterion to give the candidate a chance to demonstrate something you might have missed
• Wind down the interview: if the evidence is clear, it is fine to spend the last 15 minutes on questions the candidate has and giving them a positive experience even if the outcome will be a no

Do not give signals during the interview about outcome. Stay professional and engaged regardless of the trajectory.

Step 4 - The debrief

The debrief is where the decision is made. It should happen within 24 hours of the final interview. The longer you wait, the more evidence fades.

Who facilitates. The hiring manager, or the recruiter. Not the person with the strongest opinion. The facilitator's job is to surface evidence, not advocate.

How to prevent anchoring. Anchoring is when the first person to speak sets the direction and everyone else adjusts toward it. It produces group consensus that does not reflect individual judgements.

Prevent it with a structured debrief:

1. Everyone submits their scorecard before the debrief - no exceptions. No one reads anyone else's before they submit.
2. The facilitator reads the scores aloud at the start of the debrief. Do not share them individually beforehand.
3. Go criterion by criterion. For each criterion: the interviewer responsible shares their score and their evidence first. Then others comment.
4. The facilitator asks for evidence before discussion. "You gave a 2 on delivery - what did you see that led to that?" Not: "Does anyone agree?"

Handling split decisions. When interviewers disagree significantly (two 4s and two 2s on the same criterion), the process is:

1. Each interviewer shares their evidence
2. Look for the explanation: did they see different things? Is one piece of evidence stronger? Is one interviewer applying a different bar?
3. If one interviewer saw something concrete that others did not - weight that more heavily
4. If the split is based on different standards, go back to the scorecard anchor definitions
5. If still unresolved: the hiring manager makes the call with documented reasoning

Do not average disagreements. A 2 and a 4 do not become a 3. Understand the disagreement and resolve it.

The final decision categories:

• Strong hire: 3+ on all criteria, 4 on most, no 1s
• Hire: 3 on all criteria, no 1s
• No hire: any 1, or consistent 2s on criteria central to the role
• Strong no hire: multiple 1s, or fundamental gaps

Making the decision when the loop disagrees. If the scorecard says hire but one interviewer is strongly opposed (or vice versa), the process is:

• Ask the dissenting interviewer: "What specifically would need to be different for you to change your view?"
• If they can give evidence-based reasons - explore them
• If the objection is instinct-based ("something just doesn't feel right") - probe until there is a concrete reason or acknowledge it cannot be the deciding factor

Document the decision and the key evidence that drove it. "We hired because of X and Y, with a development area of Z" - and keep it on file.

Step 5 - Moving fast without cutting corners

Speed matters in hiring. A candidate who is strong enough for you is strong enough for everyone. The best candidates are typically off the market within 10-14 days of starting a search.

To move fast:

• Run the full loop in 5-7 days where possible. Back-to-back interviews in one or two days is better for both parties.
• Make the debrief a same-day or next-day event after the final interview - not a meeting scheduled for next week.
• Have offer approval pre-arranged before the debrief so you can move to offer within 24 hours of a positive debrief.
• Give the candidate a clear timeline at each stage: "You will hear from us by [day] with a decision."

Common delays and fixes:

Step 6 - What to do when the loop disagrees

A split result in the debrief - where interviewers have genuinely different views - is valuable information, not a problem to paper over.

If the split is on a criterion central to the role and the evidence is contradictory: lean toward no hire. You are not losing a good hire - you are avoiding a risky one. Regret about a false negative (missing a good candidate) feels different to regret about a false positive (hiring someone who does not work out and managing the consequences for 12 months).

If the split is on a secondary criterion where the evidence for a strong hire on primary criteria is overwhelming: the hiring manager makes the call, documented.

If the team says hire but the hiring manager says no: the hiring manager has the final call. But they should explain their reasoning to the team. "I heard your evidence and I weigh it differently because of X" is respectful. Just saying no without explanation is not.

If the hiring manager says hire but the team says no: this is a significant red flag. Do not override a consistent team no. Investigate the evidence. If the team is right, the manager has avoided a bad hire. If the manager believes the team is applying the wrong bar, that is a calibration problem to fix - not a reason to hire over the team's objection.

What Good Looks Like

A well-run hiring loop:

• Produces a decision within 48 hours of the final interview
• Has a documented scorecard for every candidate
• Never produces a "we liked them but we're not sure about X" outcome - the doubt is investigated or resolved
• Moves candidates through the process in 7-10 days for a full loop
• Has interviewers who can articulate specifically what they were assessing and what they saw

A well-designed scorecard:

• Has four to six criteria, each with a clear behavioural definition
• Has anchor examples of what 2, 3, and 4 look like at this level
• Is the same document used by every interviewer - no bespoke versions
• Is submitted by every interviewer before the debrief

A good debrief:

• Starts with scores shared simultaneously, not sequentially
• Goes criterion by criterion, evidence first
• Produces a clear outcome, not a hedge
• Documents the key evidence for the decision

Common Failures

No scorecard. Interviewers assess whatever they feel like. The debrief is five different people talking about five different things. No comparison is possible. Fix: scorecard before the first interview is scheduled.

Redundant interviews. Three interviewers ask behavioural questions about the same themes. The candidate is bored by interview three. Fix: design the loop so each interviewer owns a distinct set of criteria.

Briefings skipped. Interviewers walk in having read only the CV. They ask what they feel comfortable asking. The assessment is of the candidate's ability to make the interviewer feel good, not their ability to do the job. Fix: 20-minute briefing before every first interview in the loop. Non-negotiable.

The debrief is a lobby. The first person to speak strongly - usually the most senior person in the room - sets the outcome. Everyone else adjusts. Fix: scores submitted before debrief, shared simultaneously, criterion-by-criterion structure.

Slowness loses candidates. A strong candidate gets an offer from another company while you are waiting for a second debrief slot. Fix: interview slots blocked in advance, debrief scheduled before the final interview happens, offer pre-approved.

The "cultural fit" trap. Interviewers rate candidates down for "cultural fit" without being able to say what that means. This is almost always a proxy for familiarity - the candidate does not look or sound like the existing team. Fix: "cultural fit" is not a criterion. Collaboration and communication are. Score those instead.

Reference checks as a formality. References are called at the end, questions are vague, every reference says the person is great, no one updates their assessment. Fix: if you call references, use them. Ask the specific questions: "What would you do differently if you managed them again?" "What is the one area they are still developing?" "In what context did they do their best work?"

Common Biases in Hiring and Structural Counters

Checklist

Before the loop opens

• Role profile written - success at 6, 12, 24 months
• Scorecard defined - four to six criteria with behavioural anchors
• Loop stages designed - purpose, interviewer, and criteria per stage
• Interview slots blocked in diaries before sourcing begins
• Debrief slot reserved after final interview
• Offer band pre-approved

Before each interview

• Interviewer briefed on their criteria and the scorecard
• Core questions agreed for each criterion
• Candidate briefed on format and what to expect
• Scorecard template sent to interviewer

After each interview

• Scorecard submitted within two hours
• No scores shared between interviewers before the debrief

Debrief

• All scorecards submitted before the meeting starts
• Scores shared simultaneously at the start
• Criterion-by-criterion structure followed
• Evidence demanded for every score
• Split decisions resolved with evidence, not advocacy
• Clear outcome agreed - strong hire / hire / no hire / strong no hire
• Decision documented with key evidence

After the decision

• Candidate contacted within 24 hours of decision
• Offer letter prepared and sent within 24 hours of verbal offer (for hires)
• Feedback documented for no-hire decisions (in case of future appeal)
• Loop retrospective: what did the process miss or get right?

Reference: Interview Scorecard Template

Candidate:
Role:
Interviewer:
Date:
Stage: [recruiter screen / technical 1 / technical 2 / behavioural / HM screen]

Criteria assessed in this interview:
- [Criterion 1]
- [Criterion 2]

---

Criterion 1: [Name]
Score: [ ] 1  [ ] 2  [ ] 3  [ ] 4
Evidence (be specific - what did they say or do?):


---

Criterion 2: [Name]
Score: [ ] 1  [ ] 2  [ ] 3  [ ] 4
Evidence (be specific - what did they say or do?):


---

Overall impression:


Recommended outcome:
[ ] Strong hire  [ ] Hire  [ ] No hire  [ ] Strong no hire

Reasoning (one paragraph):


Questions / concerns to raise in debrief:

Reference: Debrief Facilitation Script

Open with:

"Before we start, everyone should have submitted their scorecards. I'm going to share the scores now. [Read scores aloud for each criterion.] Let's go criterion by criterion. For each one, I'll ask the primary interviewer to share their evidence first, and then we'll open it up."

For each criterion:

"[Interviewer], you scored [criterion] as a [score]. What specifically did you see that led to that?"

When there is disagreement:

"We have a [score A] and a [score B] here. [Interviewer A], what did you see? [Interviewer B], what did you see? Is there a way to reconcile these or are they genuinely different evidence?"

When instinct-based objections appear:

"I hear that you have a concern. Can you tell me what specifically you saw that drives it? I want to make sure we are deciding based on evidence."

Closing:

"Based on everything we've heard: is there any evidence we've missed or any criterion we have not discussed? [Pause.] Then my read of the room is [outcome]. Does anyone strongly disagree with that, and if so, what specifically is driving that?"