Product management
A product manager decides what to build, for which users, and how to tell if it worked. This lesson covers users, measures, trade offs and product sense. Its worked example finds out why weekly orders fell.
Firm processes and online tests change from year to year and differ by office. Use this to prepare, and confirm the exact current steps on the firm's own careers page.
Key takeaways
- A product manager owns a problem for a group of users, and a measure that shows whether it is solved.
- Write the maths of the goal.
- Use what you know about the business.
Key idea
A product manager owns a problem for a group of users, and a measure that shows whether it is solved. The five moves still work. The number at the top is a user or business measure. The facts come from how people use the product. The so what is what to build next and what to stop.
What follows describes common patterns, not any one employer's process, unless a source is named. Steps differ by company, country and level and change from year to year. Read the job advert and the employer's own careers page for the exact steps.
Product manager interviews often test a few things in separate rounds. One is how you design or improve a product for a user, often called product sense. Another is how you set and read measures, and find the cause when one moves. These are often called analytical or execution questions. Many also include estimates, sometimes technical questions, and questions about your past work. The names differ by company. Here is one example from a primary source. A post on Meta's own careers blog, dated 22 March 2023, covers a session on the first interviews. It says the session focuses on two interviews: product sense and analytical thinking. Check each company's careers page for its current steps.
The five moves, as a product manager
- 1Pin the question. Which users, which problem, and which measure says it is solved? Name one main measure and one or two guardrails, measures that must not get worse (for example, refunds or app crashes).
- 2Write the maths of the goal. Break the measure into its parts. Weekly orders are active users x orders per user. Active users are new users plus returning users. Revenue is users x conversion (the share who buy) x orders x order value. Each part is something a team can move.
- 3Use what you know about the business. How does the product make money (subscriptions, ads, a cut of each sale)? What is the company trying to do this year? That tells you which part of the tree matters.
- 4Find the facts that decide it. Split the measure by platform, country, new or returning users and date. Look for the cut where the change sits. Where you can, run a test on a share of users before deciding.
- 5Say so what. What to build or fix first, why it beats the alternatives, how you will know it worked, and what you are choosing not to do.
What is different from a consulting case
- Users come first. Interviewers want to hear which users you chose, what problem they have and why it matters to them, before any feature.
- Measures, not only money. Profit still matters, but the daily words are users, conversion, retention (how many users come back) and engagement (how much they use the product). Pair every main measure with a guardrail.
- Trade offs are the job. Engineering time is the scarce resource. Every yes delays something else, so compare value per week of build, not value alone.
- Product sense. You are expected to have views on real products: what works, what does not, and why. Before the interview, use some products and note what you would change.
Worked case
Weekly orders on a food delivery app fell: find out why
The prompt
Fictional and illustrative. You are the product manager for a food delivery app in Bengaluru, India. Weekly orders fell from 500,000 to 460,000 in one week. Active users fell from 250,000 to 230,000. Android users fell from 150,000 to 130,000 while iPhone users stayed at 100,000. A new Android version was released that week. What happened, and what do you do?
The structure
- Weekly orders = active users x orders per user
- Active users
- By platform: Android and iPhone
- New and returning
- Orders per user
- Outside causes: holidays, weather, a rival's offer
Working it through
1. Size the drop
Weekly orders went from 500,000 to 460,000. Work out the fall as a percentage.
Fall in weekly orders (percent):(500,000 - 460,000) ÷ 500,000 × 100 = 82. Orders per user before
Orders divided by active users in the week before.
Orders per user before:500,000 ÷ 250,000 = 23. Orders per user after
Orders divided by active users this week.
Orders per user after:460,000 ÷ 230,000 = 24. Where the lost users are
Android lost users while iPhone did not. Work out what share of the users lost were on Android.
Share of lost users on Android (percent):(150,000 - 130,000) ÷ (250,000 - 230,000) × 100 = 100
The recommendation
The whole 8 percent drop comes from fewer Android users opening the app. Orders per user held at 2, so the fall is not in how often people order. All of the 20,000 lost users, 100 percent, were on Android, in the same week as a new Android release. My hypothesis is that the release broke something, for example sign-in or the home screen on some phones. Today I would check crash and sign-in rates for the new version. If they confirm it, I would go back to the old version or fix the new one. Then I would watch Android users return to about 150,000. As a guardrail, I would check that the fix does not raise payment failures. I would pause the planned feature work for a day, because a broken app costs more orders than any new feature adds.
Risks: A rival's Android-only promotion could also explain the drop; Some users may not come back even after the fix.
Next steps: Compare crash and sign-in rates for the old and new Android versions; Add an alert on Android active users by app version.
Fictional. Feature A reaches 40,000 users and is expected to add 0.05 orders per user a week; it takes 2 weeks to build. Feature B reaches 10,000 users and adds 0.3 orders per user a week; it takes 6 weeks. How many extra weekly orders does feature A add for each week of build?
Using this in the interview
- Ask: which users, which measure defines success, and over what period? Is there a recent release, campaign or outside event?
- Calculate: break the measure into its parts and find which part moved. For a choice between features, work out the value per week of build.
- Say: the cause or the choice first, the evidence in one or two numbers, the guardrail you will watch, and what you will not do.
Weekly orders fell 8 percent and orders per user did not change. Where should you look next?
Why pair a main measure with a guardrail?
Sources, checked on 2026-10-02. Official pages only. Hiring steps change, so confirm them on the employer's own page before you rely on them.
- Meta Careers blog, "Product management interview Q&A: hack the initial interview", post dated 22 March 2023 (official): https://www.metacareers.com/blog/product-management-interview-qa-hack-the-initial-interview/
Sources for this lesson (2)
- Recognized public explanations of case-interview concepts and terms
- Meta Careers blog, "Product management interview Q&A: hack the initial interview", post dated 22 March 2023 (official)
My notes on this lesson
0 of 5,000 characters. Saves automatically.
Try the 3 remaining checks and drills above to complete this lesson (0 of 3 done).
Keep going: lesson 4 of 8
It builds on what you just read, in The same acumen in other roles.
Spotted something wrong or out of date? Report a mistake. We check every report and correct the page.