ChannelPulse
Four rounds, four rubrics
A PM loop is not one interview repeated. Product sense, execution, strategy and leadership score different things, and the answer that wins one round is the answer that loses another. This lesson is how to tell which room you are in, and what that room is grading.
What you'll be able to do
- Identify which of the four PM rounds you are in from the shape of the question
- Name what each round scores, and the specific failure that loses it
- Convert a vague prompt into stated assumptions and a scope you have agreed out loud
- Budget a 45-minute answer so the part that carries the score gets the time
The most expensive misunderstanding in PM interviewing is that the loop is one skill tested four times. It isn't. Four interviewers are filling in four different scorecards, and a strong answer on the wrong scorecard reads as a candidate who didn't listen.
Concretely: in a product-sense round, spending eight minutes on how you'd instrument the feature is a good instinct deployed in the wrong room — it eats the time the interviewer reserved for whether your idea is any good. In an execution round, opening with user empathy and a persona is the same mistake mirrored. Both candidates are competent. Both get "didn't quite hit the bar."
Three of the four rooms below, each with its own framework, from someone who took this loop at Airbnb, Lyft and Twitter. Note how little the frameworks share — that is the whole argument of this lesson, made concrete. The chapter at 03:28 is the closest thing to a public statement of what the scorecards contain.
Jump to the part you need
Telling the rooms apart
The prompt itself gives it away, usually in the verb.
| The prompt sounds like | You are in | It is scoring | The way people lose it |
|---|---|---|---|
| "Design X for Y." "Improve X." "What's your favourite product?" | Product sense | User empathy, structure, taste, communication | A generic feature list any candidate could have produced |
| "How would you measure X?" "Metric Y dropped 8%. Go." | Execution / analytical | Metric choice, decomposition, honest arithmetic | Naming metrics without ever defining or defending one |
| "Should we enter market X?" "Build or buy?" | Strategy | Company fit, competitive reasoning, sequencing | A framework recited with no view at the end |
| "Tell me about a time…" "Disagreed with an engineer?" | Leadership / behavioural | Ownership, influence without authority, judgement | A story with no conflict, or one where you were simply right |
Two notes on that table, because both cost people offers.
The names are not stable across companies. The same round is "product sense" at one company and "product design" or "product thinking" at another; the rubric barely moves, but the label does. If a recruiter's terminology is unfamiliar, ask what dimensions the round scores. You will nearly always get back the same short list — user empathy, structured thinking, product taste, strategic awareness, communication — and now you know the scorecard before you sit down.
Rounds are increasingly blended. Execution questions arrive inside product-sense rounds ("okay, how would you know if that worked?"), and some companies now attach a build step — a prototype produced live in the second half of a product round. Treat a blended round as a sequence of rubrics rather than a mush: finish the product argument, say explicitly "let me switch to how I'd measure this," and then switch registers. Narrating the transition is free and it tells the interviewer which scorecard to be marking.
The first four minutes decide the rest
Every one of these rounds opens with a deliberately underspecified prompt. "Design a better alarm clock" is missing the platform, the user, the business, and the constraint. The prompt is underspecified on purpose — narrowing it is part of what's being scored.
There is a bad way to narrow it, which is to ask a string of questions and wait to be told. The interviewer usually will not tell you; they will say "you decide," and now you have spent three minutes and gained nothing.
The move that works is to assume out loud and offer the exit:
"I'll assume this is a consumer mobile app, we're a mid-size company with a working payments stack, and the goal is engagement rather than new revenue. I'll also assume we're not constrained on eng headcount this half. Push back on any of those and I'll adjust."
That paragraph does four things a question would not. It shows you know which variables matter, it commits to a scope so the rest of the answer can be specific, it hands the interviewer a cheap way to steer you, and it stops the clock on the ambiguity. If they wanted a different scope they will say so within seconds, and it costs you nothing to have guessed.
Budgeting 45 minutes
Time discipline is the difference between an answer that lands and the same answer that runs out of room at the interesting part. A rough allocation for a product round:
| Phase | Share of the time | What "done" looks like |
|---|---|---|
| Clarify + assume out loud | 8% | A stated scope, and a goal in one sentence |
| User segments, and picking one | 15% | Two or three segments, one chosen for a reason |
| Pain points for that segment | 15% | Three needs, ranked, with the ranking justified |
| Solutions | 25% | Three ideas of different ambition, not three variants of one |
| Prioritise and commit | 15% | One choice, and what you are giving up |
| Measure | 12% | A primary metric plus one counter-metric |
| Trade-offs and risks | 10% | The strongest objection to your own answer, raised by you |
The two lines people underspend are picking one segment and raising the objection yourself. The first is where taste becomes visible; the second is the single cheapest way to sound senior, because it demonstrates you evaluated your idea rather than merely generated it.
The line people overspend is solutions. Three ideas is plenty. A fourth and fifth idea add almost nothing to your score and they consume the minutes you needed for the trade-off discussion — and the trade-off discussion is where the round is actually decided.
What "generic" means, and why it is the modal rejection
The most common written feedback on a competent PM interview is some paraphrase of reasonable but generic. It is worth being precise about what that means, because it is not a comment on your structure.
An answer is generic when it would be unchanged if the interviewer had named a different company, a different user, or a different year. "I'd add personalised recommendations, improve onboarding, and build a referral loop" is a fine list. It is also the answer to almost any prompt, which means it contains no information about you.
The fix is not more creativity on demand. It is specificity in three places, each of which is mechanical enough to practise:
- Name a person, not a persona. Not "busy professionals" — "someone who opens the app on the train with one hand, has 40 seconds, and will not type." Constraints generate ideas; demographics do not.
- Say the number. "That segment is maybe 4% of our base, but their usage is 3 sessions a month against an average of 12, so the headroom is where the interest is." The next lesson does this arithmetic properly.
- Name what you are giving up. Every choice has a cost. Saying it converts a list into a decision, and decisions are what the job consists of.
What to carry into the next lesson
You can now identify the room, set a scope you chose rather than waited for, and spend the minutes where the marks are. What you do not yet have is the thing that makes a product-sense answer specific: a defensible reason for the segment you picked, expressed in numbers rather than in adjectives. That is next, and it is arithmetic you can run.