Most startup founders interview by instinct. They have a conversation, ask whatever comes to mind, and then go with whoever "felt right." It's comfortable, it's fast, and it's one of the most reliable ways to hire the wrong person.

Here's the problem: unstructured interviews are barely better than a coin flip at predicting job performance. Research has shown it for decades. But we keep doing them because they feel productive — it feels like you're getting a real read on someone when you're actually just measuring how comfortable they are making small talk with a stranger.

Structured interviews fix this. They're not complicated. They don't require an HR team or an enterprise recruiting platform. You just need a clear process and the discipline to follow it.

This is that process.

What a structured interview actually is

A structured interview has three properties:

Every candidate is asked the same questions — in the same order
Questions are tied directly to the skills the job requires — not fishing for personality
Candidates are scored on the same criteria — before interviewers talk to each other

That's it. It's not about being robotic or making interviews feel like an interrogation. It's about comparing candidates on the same data instead of on vibes that differ interview to interview.

The opposite — asking different questions, following tangents, scoring based on overall impression — is what most startups do. And it produces inconsistent, bias-prone decisions.

Why gut feeling is a trap

Let's be honest about what "culture fit" usually means.

When an interviewer says "I just didn't get a good vibe," what they often mean is: "They're not like me," or "They were nervous," or "They didn't laugh at my joke." None of those things predict whether someone can do the job.

Unstructured interviews reward candidates who are good at being interviewed, not candidates who are good at the work. Articulate, confident, socially fluent people win. Quieter, more heads-down people — who might be your best engineers or operators — often don't.

Structure isn't a constraint on good judgment. It's what allows good judgment to surface.

Step 1: Build your scorecard before the first interview

This is the most important step, and most teams skip it entirely.

Before you talk to a single candidate, write down the 4-6 things a successful hire must be able to do. Not personality traits. Not "team player" or "growth mindset." Actual capabilities.

For a product manager role, that might look like:

Can translate ambiguous user problems into clear requirements
Has shipped features end-to-end with an engineering team
Can prioritize ruthlessly when resources are constrained
Communicates well in writing, with both technical and non-technical audiences
Handles conflicting stakeholder priorities without losing the thread

These become your scoring dimensions. Every interview round scores the candidate on these — and only these.

A simple 1-4 scale works well:

1 — No. Significant gap.
2 — Probably not. Some signal but not enough.
3 — Yes. Meets the bar.
4 — Strong yes. Exceeds it.

Avoid a 5-point scale. People default to 3, which makes it useless. A 4-point scale forces a decision.

If you already ran through the startup hiring checklist, you built a version of this in Phase 1. Your scorecard and your interview structure should be the same document.

Step 2: Write your questions

Now write 2-3 questions per competency from your scorecard. The goal is to give the candidate a chance to demonstrate that skill — in a way that's comparable across candidates.

There are two types of questions that work well:

Behavioral questions — "Tell me about a time when..." These ask candidates to draw on real past experience. Past behavior predicts future behavior better than hypotheticals.

"Tell me about a time when you had to ship something with an incomplete spec. What did you do?"
"Walk me through a project that didn't go as planned. What happened and what did you take away from it?"
"Describe a time when you had to push back on a stakeholder request. How did you handle it?"

Work-sample questions — Give them something close to actual work. These are the most predictive of all.

For engineers: a practical problem in their domain (not a LeetCode puzzle — something relevant to your stack or product)
For marketers: react to a piece of content or a campaign brief
For designers: walk through a past project and explain the decisions
For operations: here's a messy process, how would you approach untangling it?

What to avoid: puzzle questions ("How many piano tuners are in Chicago?"), gotcha questions, and anything that's really just testing how relaxed someone is. These feel like signal. They're mostly noise.

Step 3: Standardize your rounds

For most startup roles, three rounds is the right number. More than that and you're burning everyone's time. Less than that and you're flying blind.

Here's a structure that works:

Round 1: Intro call (20-30 min)

Mutual fit check. You're not going deep on skills yet. You want to know:

Can they do this type of job based on their history?
Do they understand and want this specific role at this specific company?
Are compensation, location, and timing in the right ballpark?

Ask 2-3 light behavioral questions. Keep most of the time for their questions about the role. Candidates who ask good questions are usually better candidates.

Round 2: Skills deep-dive (45-60 min)

This is the core round. Go through your behavioral and work-sample questions for the competencies that matter most. The same questions, in the same order, with every candidate.

One interviewer leads, but a second can observe and take notes. Rotate who leads if you have multiple interviewers on this round — just keep the questions consistent.

Round 3: Team conversation (30-45 min)

Introduce them to someone they'd work with closely — not a senior leader, the actual peer. Two goals here:

Let both sides get a realistic preview of the working relationship
Get a perspective from someone other than the hiring manager

This isn't a vibe check. Give your team member 2-3 specific questions to ask and a simple scoring sheet. Otherwise, "they seemed great" tells you nothing.

Step 4: Score independently, then debrief

After each interview, every interviewer scores the candidate on your 4-6 criteria — before talking to anyone else. Write it down immediately. Memory decays fast, and you'll be comparing candidates weeks later.

Then debrief as a team. Share scores. Surface disagreements. The goal of the debrief isn't to reach consensus — it's to understand why there's disagreement.

Run your debrief with one rule: specific observations only.

"She struggled to explain her technical decisions to a non-technical audience" → useful
"I didn't feel like she was that excited" → not useful
"He'd already solved a nearly identical problem at his last company" → useful
"He seemed a bit stiff" → not useful

Impressions are conclusions without evidence. Observations are evidence. Require evidence.

Common mistakes to avoid

Writing questions after the interview. If you think of a great question during the interview and only ask one candidate, you can't compare that answer to anyone else's. Save it for the next cycle.

Letting one strong answer carry a weak overall score. A candidate who absolutely nails one competency but misses three others is still missing three others. Average the scorecard, don't cherry-pick.

Interviewing too many people. If you're doing more than three rounds, you've lost confidence in the process — and you're signaling it to candidates. Define your bar upfront and trust it.

Skipping the debrief. It's easy to skip when the hiring manager has already made up their mind. Don't. The debrief is where hidden bias gets surfaced, where a detail one interviewer caught gets shared with the rest of the team. It's the final check before a decision that's hard to reverse.

Treating round 3 as optional. "We really liked them in round 2, let's just move to offer." You're cutting the team out of a decision that affects them. Run all three rounds.

A template to steal

Here's a condensed version you can adapt for almost any role:

Scoring criteria (customize per role):

Core skill 1: _____________
Core skill 2: _____________
Communication (written/verbal): _____________
Problem-solving under constraints: _____________
Culture/working style alignment: _____________

Behavioral questions (ask all, in order):

Walk me through a project you're most proud of and why.
Tell me about a time when something you owned didn't go as planned. What happened?
Describe a situation where you had to influence someone without formal authority.
What's the hardest feedback you've ever received, and what did you do with it?

Work-sample prompt: (Design this specifically for the role. Make it something you'd actually ask them to do on day 10.)

Score after each round (1-4 per criterion): | Criterion | Score | Evidence | |-----------|-------|----------| | Core skill 1 | | | | Core skill 2 | | | | Communication | | | | Problem-solving | | | | Working style | | |

Customize the criteria and the work-sample prompt. Keep the behavioral questions consistent across all candidates.

The bottom line

A structured interview doesn't guarantee you'll hire the right person every time. But it dramatically increases your odds — and it protects you from the most common traps: hiring people who interview well, passing on people who just needed more time to warm up, and making decisions on inconsistent data.

You don't need an HR team for this. You need a scorecard, a set of questions, and the discipline to ask them the same way every time.

The companies that hire consistently well aren't the ones with the best instincts. They're the ones with the best process.

Already running a structured process? Make sure candidates have a great experience before the interview too. Set up your career page with Pipol — free, powered by Notion, live in minutes.

How to Run a Structured Interview Without an HR Team