Most Interview Processes Are Broken. Here's How to Fix Yours.
The average interview process is designed to protect the company, not to find the best person. Here's what a better process looks like.

I’ve sat in a lot of interview rooms. On both sides of the table. And the thing I’ve noticed is that most processes are optimised for the wrong thing, minimising risk for the company rather than genuinely understanding the person in front of you.
The result is predictable: bloated processes that drag on too long, generic questions that reward rehearsed answers, and decisions made on gut feel dressed up as structured evaluation. You end up hiring the best interviewer, not the best candidate.
Here’s what a better process looks like.
Match the depth of the process to the stakes of the role
The first mistake most teams make is applying the same process to every hire, regardless of seniority. An intern and a Lead Developer are not the same decision. Treat them the same and you’ll either over-engineer the junior hire or under-scrutinise the senior one.
The right number of conversations depends on the level of responsibility and the long-term impact that person will have on the team. Here’s a simple framework:
Four is the ceiling. If you haven’t gathered enough signal in four conversations, more interviews won’t save you; they’ll exhaust the candidate and signal that your team can’t make decisions.
Four lenses every interview should cover
Regardless of the role or the number of sessions, a good interview process should consistently cover the same four areas. Not as a checklist, as genuine lenses on who the person is and how they work.
The best interview processes don’t just evaluate candidates. They give candidates a genuine reason to say yes.
Stop running the same technical test you ran in 2015
Generative AI has fundamentally changed what a take-home test can tell you. A polished solution submitted 24 hours later is evidence of almost nothing. Teams that haven’t updated their technical evaluation since the pre-AI era are essentially testing how well someone can prompt a language model.
The fix is to move toward observed, collaborative formats. Pair programming. Live code review. Architecture discussions. These can’t be outsourced to a tool, and they reveal how someone actually thinks, not just what they can produce.
One practical rule: tell candidates in advance what tools they’re allowed to use. That’s not hand-holding, it’s respecting their time and getting a more honest picture of how they work.
Consistency isn’t the enemy of judgment; it’s what makes judgment trustworthy
I’ve heard hiring managers resist structured processes because they feel constraining. That’s the wrong framing. Consistency isn’t about removing human judgment from hiring; it’s about making sure that judgment is applied to the same criteria, not to whoever the interviewer happened to warm to that day.
Evaluate every candidate against the same dimensions. Then bring your instinct to bear on those dimensions, not instead of them. And before any process begins, have it reviewed by the person with the most context on where the team is going. That conversation will surface assumptions you didn’t know you were making.
What you do after matters as much as what you do during
Every candidate who goes through your process forms an opinion of your company. The ones you don’t hire talk to people. The ones you do hire remember how they were treated before they accepted.
Give everyone timely, honest feedback. If you’re moving forward, be clear about what comes next. If you’re not, say so, and say why, where you can. That’s not just courtesy. It’s a signal to the market about the kind of company you are.
The interview process is the first chapter of a relationship. Make it one worth continuing.
The question worth sitting with: if you removed every candidate’s name and photo from your last ten hires, would the pattern of decisions still hold up?
Written as part of an ongoing series on engineering leadership and team building.
- Written by
- Federico Corradi
- Published
- May 18, 2026
- Reading time
- 4 min read
- Topics
- Leadership, Thought Leadership
- Edition
- N° 002 / 2026
