Companies and interview loops
What happens in a PM interview loop, round by round
In this lesson9 sections
An unofficial lesson drawing on the methods in the Behind the Interview Scenes chapter of Cracking the PM Interview by Gayle Laakmann McDowell and Jackie Bavaro (2014). PracHub is not affiliated with the authors or publisher. Buy the book
A PM interview loop is a chain of rounds, and each round is there to collect one kind of evidence about you. Once you know which round wants which evidence, you can prepare for each room on its own terms instead of for "the PM interview" in general.
We'll follow one invented candidate, Maya, from the first call to the last round. She has four years in product, the last two as a PM at a payroll software company, and she's applying to the booking team of a mid-size marketplace where people rent event venues, such as private dining rooms and rooftops, for parties and work events. The marketplace earns a commission on every booking made through it.
The template under almost every PM loop
Take the company names away and most PM hiring runs in four stages:
A recruiter screen, often 20 to 30 minutes.
One or two screens with a PM or the hiring manager, usually on video.
The loop: usually four to six interviews, often on a single day, and often virtual.
A decision, made by a hiring committee, by the hiring manager, or by the interviewers together in a debrief.
Learn this template first, then learn how each company bends it. The bends catch people out: a written exercise between stages, say, or a decision made by people you never meet. The last stage has its own lesson, how PM hiring decisions are made.
Treat any description of a company's process, including this course's, as a guess until someone current confirms it. Cracking the PM Interview profiled several companies' processes as they ran in 2014, and most of those details have changed.
Maya's first source on this marketplace's loop was a candidate's write-up from a year earlier, which mentioned nothing written. Before she planned around it, she asked her recruiter, and a PM who had joined the booking team that spring:
"Before I start preparing, could you walk me through each round: who I'll meet, what each person covers, and whether there's anything written or take-home?"
The recruiter's prep document listed a take-home memo, added since that write-up. When someone describes a loop, ask how long ago they went through it. Many recruiters keep a prep document; if yours doesn't offer one, ask.
The recruiter screen is a real interview
Recruiters turn people away at this step, and their questions have better and worse answers: why this company, why leave your job now, what scope you've had, what level you're aiming for and what pay you expect. Some add a short product question to hear how you think out loud.
Maya prepared for the call as she would for a PM interview. She timed a two-minute introduction and a three-sentence answer to why this company, and rehearsed both aloud. She also wrote one honest sentence on why she was leaving, since a vague answer there makes a recruiter wonder what you're leaving out. On pay, she planned to ask for the level's range before naming any number.
She kept the last ten minutes for the question above. PracHub's guide to the PM recruiter screen covers the common questions.
To practise this step, try Answer first-round HR questions. Answer each part aloud, in the minute or two a recruiter would allow.
A screen tests range in one short session
The screen with a PM or the hiring manager is often broader than any single loop round. The interviewer has about 45 minutes to judge whether you merit a full loop, so they test range rather than depth in one area. Some companies also hold a separate technical screen, often with an engineer, before the loop; your recruiter can confirm it.
Maya's screen would pair a product question with a metric one. The weak spot in that pairing is the switch: spend too long on the idea and the measure gets squeezed. So she practised going straight from an idea to how she'd measure it, in one answer. Know the company's main product as a regular user would, so your ideas start from the real thing.
To practise this step, try Favorite Products & Improvement Metrics, which asks for an improvement and then the metrics that would show it worked, much as a screen can.
In the loop, each interviewer writes up one area
Product sense, analytical thinking, leadership and drive (how you work with people and push work through), technical understanding: most loops share these out among the interviewers before you arrive, and each person's written feedback covers their own area first. The names vary by company. Some loops add strategy, or a round on building products with AI.
Some interviewers play a fixed role instead, such as a partner from another function judging whether their team could work with you. Knowing what each person is there to find out tells you what to show them.
So ask the recruiter who covers what. Maya's loop had five interviewers, and only the hiring manager was a PM:
| Round | Interviewer | Area they write up | What Maya prepared |
|---|---|---|---|
| 1 | Data scientist | Analytical thinking | A metric, a guardrail and a test for one booking change |
| 2 | Design lead | Product sense | The booking flow seen from an organizer's side, and its trade-offs |
| 3 | Support lead | Working with partners, customer focus | A story where support tickets changed her plans, and how she readies support for a launch |
| 4 | Engineering manager | Technical understanding, scoping | How booking works behind the screen, and a story of cutting scope |
| 5 | Hiring manager | Leadership and drive, motivation | Her two strongest impact stories, and why this team |
The table changed her plan: only one round in five was mainly product sense, so she moved hours from product design into the analytical round and into stories about partners.
Change how deep you go, and which words you use, for each person. An executive wants the decision and its business effect first, with detail on request. The engineering manager wants to hear that you know what's hard to build and what can wait. The design lead cares where an organizer hesitates, not about sample sizes. Maya had one payroll launch story in three versions: for an executive, the result in one number; for the engineering manager, the dependency she dropped to ship on time; for the designer, what users misread in the first release.
For the product sense round, try Choose the first Dropbox feature to build for distributed teams, and defend the pick, and answer it as you would to a design lead.
Partners in the loop, such as a support lead or a tech lead, are also asking a quieter question: would I want this person around when something breaks? Show them by how you talk about their work. Maya's support story gave that team its credit:
"Our support team tagged about 300 tickets in one month about direct deposits landing a day late. Their tags traced it to a single bank, and we fixed it in two weeks. Now I send support the launch notes a week ahead and ask what customers will ask."
Analytical and technical rounds
The analytical round often asks for two things in one answer: the metric that would show a change worked, and how you'd prove the change caused it. Practise the two together, starting from what the user would do differently if the change worked.
Maya practised on her own team's problem. Venue listings show only the room fee, and the full cost (minimum spend, service charge, cleaning) arrives later in a quote. The team wants to show an estimated total in search results. Her goal: organizers who now walk away from a surprise quote would book a venue they can afford instead. Her main metric, which rises if it works: confirmed bookings per search. Her guardrail: quote requests per venue, held above an agreed floor, because showing higher prices up front will cut requests by itself, and venues judge the marketplace by its leads. For the test, she'd split by city rather than by organizer, since organizers in one city compete for the same venues and dates. Metrics and diagnosing a drop covers the metric side in depth.
To practise this step, try North Star Metrics & Experiment Design, which asks for a metric and an experiment in one answer, and Choose one improvement for X (Twitter) and define the metrics that prove it worked.
North Star Metrics & Experiment Design
Practice defining a North Star Metric and experiment for a Microsoft product using Teams as the example.
Choose one improvement for X (Twitter) and define the metrics that prove it worked
A product sense question asking how you would improve Twitter, now X, and which success metrics would show the change worked.
Some companies lean hard on estimation and technical questions drawn from their own business: a commission marketplace asks you to size bookings, a subscription company asks about subscribers and cancellations. So practise estimation on the company's main source of revenue.
For Maya's marketplace, that's bookings × average booking value × commission. Her drill, with every input her own assumption: a city of 1,000,000 people has about 400,000 households. Say 6% of them rent a venue for a private event each year: 24,000 events, or 2,000 a month. If a quarter of those go through the app, that's 500 bookings. At $2,000 a booking and a 10% commission, each earns $200, so private events bring in 500 × $200 = $100,000 a month. The input she trusts least is the quarter booked through the app, so she'd check it from the venue side: how many venues in that city list on the app, and how many evenings a month does each fill through it? Our lesson on market sizing building blocks has more.
Technical rounds for PMs are rarely coding tests now; most ask how a system works and which trade-offs you'd accept. Maya prepared to explain what happens when an organizer taps "Book": the app checks the venue's calendar, holds the date, authorizes the deposit and notifies the venue. Then one trade-off, out loud. Holding a date for ten minutes while someone pays prevents double bookings, but another organizer ready to pay at once sees the date as taken and may book elsewhere. Confirming only after payment keeps dates open, and now and then two organizers pay for the same evening. Our lesson on technical rounds for PMs goes further.
Take-home and written exercises
Some loops add a written piece: a memo or a short deck prepared at home, or a timed writing exercise. PracHub's guide to the Amazon writing assessment shows one version. Whoever reads it is hunting for your decision and the evidence for it, and each sentence that carries neither makes the hunt longer. Put the recommendation in the first two lines.
Maya's take-home asked whether organizers should be able to cancel for free until shortly before the event. Her first draft opened with two paragraphs of market background. She moved the answer to the top:
"Recommendation: offer free cancellation until 30 days before the event on bookings under $3,000. I expect the share of searches that end in a booking to rise from 4% to 5%. The risk is venues losing evenings they could have sold, and I'd end the offer if more than 8% of these bookings cancel inside the free period."
Evidence, risks and her first test followed, each paragraph opening with its point. She cut adjectives and kept sentences short. She read the memo aloud to catch typos, then gave it to a friend who knew nothing about the company and rewrote every sentence the friend had to read twice.
To practise this step, treat Go-to-market plan and launch deck for Google's web analytics product as a real take-home: give yourself two hours, and make the recommendation your first slide.
When each team hires for itself
Pool hiring, common in new-grad and associate PM programs, judges you on general PM skill and matches you to a team after the hiring decision, so prepare evenly across every area. Under team hiring, the hiring manager's needs shape the loop. The booking team's job post asked for someone who weighs a change's effect on organizers and venues alike. The same marketplace's pricing team, which had turned Maya down eight months earlier, had wanted someone who had run pricing experiments. So read the team's post, not the company's general page about PMs. Expect some interviewers from outside the team, though, judging you against the company's general bar, so keep your range.
A no from one team often means you didn't match that team's needs that quarter, not that the whole company has judged you. Maya's recruiter confirmed the booking team didn't hold the pricing rejection against her, since the two teams hire separately. If a team says no, ask whether other teams are hiring, and whether a waiting period applies before you apply again.
Referrals
A referral usually gets your application read sooner, and by a person. Some companies fill a large share of roles this way, and someone who works there can tell you whether yours does. At those companies a warm referral beats a polished cold application. Warm means the person referring you has worked with you and can say so in their note.
Maya's cold application to the pricing team had waited five weeks for a reply. This time she wrote to a former payroll teammate, an engineer now at the marketplace, with the role's link, her resume and two sentences on why she fit that he could paste into the referral form. He added a line about a launch they had shipped together, and the recruiter called within a week.
A checklist before any company's loop
Who makes the decision: a committee reading written feedback, the hiring manager, or the interviewers in a debrief?
Are you hired by one team, or into a pool that's matched to a team later?
How many rounds, how long is each, and are they on video or in person?
Which area does each interviewer cover, and which area does this company weigh most?
Is there a take-home, a writing exercise or a presentation?
Does the company lean on estimation or technical questions, and about which part of its business?
What has changed since the most recent report you read?
Answer them from current sources: the recruiter's prep document first, then recent candidate reports such as the PM questions reported at Google, Meta and Amazon, and company guides such as PracHub's Google and Amazon PM interview guides. For the product and market side, see researching a company before a PM interview.
Then rehearse the loop in its real shape. Maya asked two friends to run a mock day of five rounds back to back. They split the five interviewers from her table between them and graded each round only on that interviewer's area. Our guide to running a PM mock interview covers the questions, follow-ups and debrief.