How to hire a product manager
Most PM hiring loops test presentation skill and produce confident mis-hires. A four-stage process that tests judgement instead, the scorecard to run it with, and the false negatives it will save you from.
The standard product manager interview loop is four conversations, a case study, and a decision made mostly on how articulate someone was in the room. It reliably selects for people who interview well, which correlates with the job less than anyone would like.
The fix isn’t more stages. It’s testing the thing the role actually consists of: making a defensible decision with incomplete information, and getting other people to act on it.
Start by writing the role as a decision, not a list
Most PM job descriptions are interchangeable. “Own the roadmap, work cross-functionally, drive outcomes” describes every product job that has ever existed, which means it filters nobody and attracts everybody.
Replace the responsibilities list with three sentences:
- The decision this person owns. “You decide what we build in the onboarding and activation area, and you can say no to sales.”
- The problem they inherit. “Activation is 30% and we don’t know why. Two previous attempts failed.”
- What success looks like in a year, stated as a number or a state, not an activity.
This is uncomfortable to write because it forces you to admit what’s broken. It also does more filtering than any screening call: the candidates who respond to that ad are self-selecting on the actual problem.
If you can’t write sentence 1 — if nobody in your company can genuinely say no — stop and fix that first. Hiring a PM into an environment where no one is allowed to decide anything produces a coordinator with a PM title, and they’ll leave inside a year.
The four stages
1. Screening (30 min) — is there a real track record?
One question does most of the work: walk me through a decision you made that your manager disagreed with.
Candidates without independent judgement can’t answer it. Not because they’re weak, but because they’ve never been given the room, and you’ll find that out in thirty minutes rather than in month three.
2. The work sample — on your actual problem, not a puzzle
Give them a real, current, non-confidential problem from your product. A page of context: the data you have, the constraint, the stakeholder positions. Ask for one page back plus a 45-minute conversation.
What to give them: something you genuinely haven’t solved. It’s tempting to use a problem you already answered so you can compare — resist it, because you’ll score for agreement with your own conclusion rather than for quality of reasoning.
Pay for it if it’s over two hours. Long unpaid exercises select for candidates who aren’t currently busy, which is not the filter you want.
What you’re reading for, in order:
- Did they ask for information before answering? (The best candidates come back with questions first.)
- Did they name what they’d do and what they’re giving up?
- Did they state a condition that would change their mind?
- Would an engineer be able to start from this?
3. The cross-functional conversation — can they be disagreed with?
An engineer and a designer, together, pushing back on the work sample. Brief them to actually disagree, not to be pleasant.
You’re watching for one thing: does the candidate update, hold, or crumble? All three happen and only one is bad. A candidate who changes position under mild pressure will change it under sales pressure too. A candidate who never updates will be exhausting. The good answer looks like “that’s a better point than mine about X, I’d still hold on Y, and here’s why.”
4. The reference call you actually make
Two questions, to a former peer rather than a former manager:
- “What did they decide that you disagreed with, and how did they handle it?”
- “What would you not put them in charge of?”
The second question gets real answers surprisingly often, because it’s framed as fit rather than criticism.
The scorecard
Score the four things that predict the job, independently, before the panel talks to each other. Discussing first is how one confident voice becomes a consensus.
| Dimension | Evidence you’re looking for |
|---|---|
| Judgement | Named the trade-off, not just the choice |
| Communication | An engineer could start from their page |
| Influence | Held or updated a position under real pressure |
| Curiosity | Asked for information before answering |
Deliberately absent: culture fit and enthusiasm. Both are real, both are also where bias enters most easily, and neither predicts performance.
The questions that separate senior candidates go deeper on what a strong answer sounds like in each of these — worth reading from the other side of the table before you write your loop.
The false negatives this saves you from
Three groups that good candidates fall into and standard loops reject:
People from a company with no product culture. They’ll have no framework vocabulary and often excellent instincts. Test the instincts.
Career changers with domain depth. A support lead who’s spent three years hearing exactly where customers get stuck knows things no external hire will learn in six months. They interview badly against PM-shaped candidates and frequently outperform them.
Quiet candidates. The loop above rewards written reasoning over verbal performance, which is the whole point. Some of the best PMs I’ve worked with were unremarkable in a room and devastating on a page.
After they join
The hire isn’t done at the offer. Two things in the first 90 days decide whether it works:
- Let them make a decision in month one and back it publicly, even a small one, even if you’d have chosen differently. The organisation is watching to learn whether this person’s decisions are real.
- Give them one thing to kill. Nothing establishes a product manager’s authority faster than stopping something that shouldn’t be running — and nothing undermines it faster than being overruled the first time they try.
If you can’t do either of those yet, the honest read is that the function isn’t ready for the hire — which is a different problem, and the one worth diagnosing first.
If you’re hiring into a gap you’re currently filling yourself, the search and the vacancy are two separate costs. Interim cover is one way to stop paying the second one while you take your time over the first — and the interim usually ends up writing a better role description than the one you’d have started with.