Picture this. You are a Product Manager at Netflix. You have barely settled into your chair when the interviewer slides one sentence across the table: “For the past few weeks, your average weekly watch time has been going down. How would you approach this problem?”
That is it. No data, no context, no hint. Just you, a blank page, and a ticking clock.
This is the case study round, and in my experience it is where most Product Manager offers are quietly won or lost. You can have a great resume and a warm chat in the first ten minutes, but this is the round where the interviewer finds out whether you actually think like a PM. So let’s solve one together, end to end — the exact Netflix “drop in watch time” case I walked through in this mock interview.
Why the case study round decides your PM interview
A case study is not a quiz with a right answer at the back of the book. It is a conversation. The interviewer is not sitting there waiting for you to magically name the one bug that broke watch time — they are watching how you get there.
Can you take a vague, scary, open-ended problem and break it into pieces? Do you reason out loud so they can follow your thinking? Do you know when to ask for data and when to make a call? That is the job. The interview is just a compressed, slightly terrifying version of a normal Tuesday.
So the goal is not to be right. The goal is to be followable.
The mistake most candidates make in the first two minutes
Here is what I see go wrong more than anything else in the mock interviews I run: the moment the candidate hears “watch time is dropping,” their brain sprints to solutions. “Let’s send a re-engagement email! Let’s improve recommendations! Let’s add more content!”
Please don’t. You have diagnosed nothing. You are prescribing medicine before you have asked where it hurts.
A strong candidate does the opposite. They slow down. They treat the first few minutes as detective work, not firefighting. Watch how the whole case below is basically one long, patient narrowing-down — and how the actual solutions only show up right at the end, once we know exactly what we are fixing.
Step 1: Clarify the problem before you touch it
Before anything else, make sure you and the interviewer are solving the same problem.
I started by pinning down the metric itself: watch time, for me, is the total number of minutes of content watched across all users, including rewatches, across all devices and geographies. The interviewer agreed and added a useful constraint — keep it at a weekly level to avoid daily noise. Good. Now “watch time” means the same thing to both of us.
Then I scoped it: is this drop global, or a specific market? The interviewer narrowed it to India only.
Notice that these are not throwaway questions. Defining the metric and scoping the market took thirty seconds and instantly made the problem smaller and sharper. That is the whole game — turning one giant fuzzy problem into a series of smaller, answerable ones.
Step 2: Split the funnel into “internal vs external”
Now the detective work. Before proposing anything, I need the root cause, and for that I need data.
I asked for three things: is the drop on a specific platform (app vs web), how many users are logging in, and what is the average watch time per user? The answers did most of the narrowing for me:
- The drop is across all platforms — so it is not a broken app build.
- The number of users is stable — same traffic, roughly the same split of new versus returning.
- But average watch time per user is down about 10%.
Sit with that for a second, because it is the hinge of the entire case. Same number of people are showing up, but each one is watching less. That single fact tells me the problem is almost certainly internal — something about the content or user behaviour — not external.
Now, here is a habit that separates strong candidates from average ones: I still listed the external factors before discarding them. Could it be a competitor launch? A change in our marketing? A cricket World Cup stealing everyone’s evenings? I said each one out loud, explained why the stable user base makes it unlikely, and only then set them aside.
Why bother, if I was going to reject them anyway? Because if you silently skip a factor, the interviewer can’t tell whether you rejected it on purpose or never thought of it at all. Say it, reason about it, discard it. Out loud.
Step 3: Follow the trail to the real culprit
With external factors parked, I went hunting inside. Two suspects: content and user behaviour.
I started with content. Did we pull anything off the platform? Is the drop in regional or international content? Any particular genre? The interviewer’s answers kept sharpening the picture: roughly 70% of what India watches is regional — mostly Hindi — and the majority of the drop is coming from that Indian content, with no genre pattern. Supply and quality of that content, they told me, are fine.
So the story now reads: same users, watching less Indian content, and nothing wrong with the content itself. That points squarely at user behaviour.
I probed the obvious behavioural lever first — did the search or recommendation algorithm change? No. Then I asked whether it was tied to a content type, and got a gift of a clue: a higher drop in movies, especially old movies. Stitch it together and the problem has a face now — it is users who used to watch old Hindi movies.
But I still needed to know how those users used to reach that content. So I listed behavioural patterns and asked the interviewer to point me in the right direction: time of day, online versus offline downloads, and how they discovered content — home-page thumbnails, notifications, or search. Timing was too volatile to matter. Online/offline hadn’t shifted. But search? A significant chunk of the drop was coming from users who used to search for content and watch it.
We have now travelled from “watch time is dropping globally” all the way to a single sentence: users who used to find old Hindi movies by searching for them are watching less. That is the power of narrowing before solving.
Step 4: Break the journey into a funnel, then find the break
When the drop is tied to search, walk the search journey as a funnel. There are only a few places it can break:
- Users aren’t clicking the search icon at all.
- Users click search, but don’t actually type anything.
- Users search, but the results disappoint them, so they drop off.
I flagged that number three points at the search algorithm — which the interviewer had already ruled out. That left one and two, and sure enough, both were down: fewer people clicking search, and fewer people typing a query after clicking.
So I asked the obvious follow-up: did the search experience change recently? It had. Two changes, in fact. The search icon had moved from the bottom nav bar to the top right, and the search page now surfaced more recommendations, including Games.
There it is. Not a mysterious algorithm, not a content drought — a UI change that quietly made an old habit harder.
Step 5: Solve it without the easy button
Now the interesting twist, and a very realistic one. The obvious fix is “just roll back the change.” But the interviewer took that off the table — those changes support a longer-term strategy (pushing the Games feature), and Netflix prefers not to keep flip-flopping the design. So: fix the watch-time drop without reverting. This is where a lot of candidates freeze. Constraints are a feature of the job, not a trick.
Before suggesting anything, I did something I’d tell every candidate to do — I opened the actual Netflix app to experience the search journey first-hand. Then I proposed fixes mapped to each broken step:
- For users who click search but don’t type: clean up the search page. When someone taps search, they usually already have something in mind; a page cluttered with recommendations (and Games) is a distraction that pulls them away and loses the session. I hypothesised we get little real watch time from those search-page recommendations — worth validating with data — and if so, strip them back for a test cohort.
- For users who don’t click search at all: education. The icon moved, so a few sessions of coach marks or highlighters make sure people notice its new home. Basic, but easy to under-do.
- For users who log in and leave without watching anything: an exit nudge — “Can’t find what you’re looking for? Try our new and improved search.” — to pull them back to the very feature that changed. (The interviewer rightly noted its limits: you can catch a back-button exit, but not someone hitting the home button.)
- For users still scrolling the home feed: introduce a search prompt inside that flow — a search bar or a nudge after the first or second scroll — so the new layout gets discovered where people already are.
Every one of these is a testable experiment on a small cohort, not a bet-the-company redesign. That framing — hypothesis, small test, measure — is what “product sense” actually sounds like.
The curveball at the end: why only Hindi content?
Just when I thought we were done, the interviewer threw one last question: the search change affected everyone, so why did it hurt only Hindi content and not English?
This is the kind of question that looks like a trap and is really an invitation to show depth. My reasoning: the drop is concentrated in the tail — old movies. Tail content rarely surfaces in recommendations, so the only practical way to reach it is search. And in India, demand for that tail — classic old Bollywood — is far higher than demand for old international titles. So when search got harder, Hindi tail content took the hit while English barely moved, because search was never driving much English watch time in India anyway. Easy enough to verify by comparing search’s contribution to old international content before and after the change.
You will not always get this question. But when you do, resist the urge to hand-wave. A clean, logical explanation here is often the difference between “good candidate” and “let’s make an offer.”
The five things that actually get you hired
Strip away the Netflix specifics and here is what this case is really teaching. These are the habits I look for in every mock:
- Ask questions until you truly understand the problem. Define the metric, scope the market. Thirty seconds of clarifying saves you from solving the wrong thing for twenty minutes.
- Be proactive. Offer your own definitions, assumptions and ideas, and ask the interviewer to confirm — don’t sit waiting to be spoon-fed.
- Don’t be afraid to ask for help. Asking for the right data point is not weakness; it is exactly how the job works.
- Be explicit, even when you reject an idea. Say why you’re discarding a factor. Otherwise the interviewer can’t tell the difference between a considered rejection and a blind spot.
- Take your time. You are a human, not a supercomputer. Pause, structure your thoughts, and don’t blurt out the first idea that pops into your head.
Reread the walk-through with those five in mind and you’ll spot every one of them in action.
Frequently Asked Questions
What is a product manager case study interview?
It’s a round where the interviewer gives you an open-ended product problem — like a drop in a key metric, or “design an app for X” — and watches how you break it down, ask for data, and reason towards a recommendation out loud. It tests structured thinking and communication far more than domain trivia.
How long should a PM case study answer take, and is there a right answer?
Most cases run 20–30 minutes, and there usually isn’t a single right answer. Interviewers care about the journey: did you scope the problem, narrow logically, weigh trade-offs, and stay easy to follow? A well-reasoned “here’s where I’d dig next” beats a lucky guess with no structure.
How do I get better at PM case study interviews?
Practise out loud, with someone playing the interviewer and pushing back on your assumptions — reading frameworks silently isn’t enough. Time yourself, narrate your thinking, and get feedback on where your logic jumped. That feedback loop is the whole point of a mock.
Case studies feel intimidating because they’re open-ended — but as you just saw, “open-ended” is exactly what you can turn into a calm, structured conversation with a bit of practice.
If you’d like to run one of these live and get honest feedback on where your thinking is strong and where it wobbles, that’s exactly what I do. I’ve mentored 50+ people into and up through PM roles at Google, Microsoft and Amazon. Book a mock interview and let’s practise for real.
And if you’re earlier in the journey, start with how to break into Product Management and the most commonly asked PM interview questions.










