The Ultimate Guide to Product Sense Interviews
How to crack the No.1 product interview question. An actionable guide that is relevant in 2024.
In one of my previous articles, I covered the interview process at the top 25 global tech companies.
One thing stood out: the most frequently asked interview question by far was product sense. “Design me an X for the task of Y” accounted for 58% of all the questions.
There’s almost a guarantee you’ll encounter it when interviewing at any mid-to-large cap product company.
Aakash Gupta goes as far as saying, “It’s a red flag for the hiring process or decision-making culture if you’re not asked one.”
Why is it so prevalent?
Firstly, it offers valuable insight into the candidate’s thinking process.
Is the potential hire a “backlog operator” or someone who can empathize with users and map their problems systematically?
Secondly, as Kevin Yien succinctly noted on a recent Lenny’s podcast:
Product Sense is a skill of making decisions without data
In essence, this question simulates product work on a smaller scale.
According to internal data from Meta and Google, the results of this interview are one of the top predictors of a candidate’s success in the role.
Given its significance, if you’re interviewing for a product position, mastering the product sense case is essential.
This guide will show you how.
It is based on my personal experience as an interviewer (with 500+ interviews and counting) and intel from the market.
Today’s article
The why behind the product sense. What is it, which skills are evaluated and how, types of questions.
The perfect response. How to structure your ideal response.
Advanced tips. How to go beyond what’s expected.
📝 Word count: 2627 words
⏱️ Reading time: ~19 minutes
🤔 The why behind the product sense
🗂️ Types of questions
Product sense comes in three types, differing by the amount of context provided.
Launching a new product comes with the highest ambiguity and limited context.
On the other hand, launching a new feature already comes with the context of an existing product and reduced uncertainty.
New product
This one is the most straightforward. It typically comes as “Design X to do Y for the Z persona”.
Design a trip organizing app,
Design an offline tour app for cities,
Design a dashboard to be used by police officers to record crimes,
Design a vending machine for blind people,
Design a web-based application for collaborative work,
Design a Facebook product for learning how to play a musical instrument,
Build a product for hyperlocal communities,
Design a microwave for college students,
Redesign the Netflix experience solely for seniors
New feature
Fundamentally similar, but the difference lies in context. For a new product, you must build the fundamentals from scratch. For a new feature, you reverse-engineer and map the user base of an existing product.
How to determine whether there’s a need for a lending feature in the marketplace,
Design a parking solution for Google Maps,
You are the PM working for the payments team. How would you introduce a crypto-currency as a new method of payment?
Improve existing product
Similar to the new feature, but with full autonomy for you to make choices on what to focus on.
Choose a product that’s not popular, how would you improve it?
How would you improve Spotify?
What’s your favorite product?
How would you solve for the biggest complaints in Turo
❓ What is it for?
I love giving product sense interviews. It allows me to evaluate a diverse set of skills in a really limited period of time (~30 minutes).
All the answers are reflective of whether or not the candidate will succeed in a role.
What I’m typically looking for:
Systematic thinking
Is the candidate’s thinking all over the place, or does she break large and ambiguous problems into smaller and interconnected chunks?
Candidates who do not solve problems systematically will face a bunch of problems downstream.
First, they won’t be able to decouple the signal from the noise and build a proper strategy. As a consequence, they will be reactive and focus on pleasing their stakeholders by shipping “important features.”
Secondly, they will have a hard time drafting a compelling narrative and pitching it to the stakeholders and peers.
Thirdly, on an execution level, they will fall for the first solution at hand instead of systematically going through each and making a trade-off.
You can learn more about developing this skill in one of my previous articles.
Strategy and decision making
Can a candidate make decisions by herself or does she require clear guidance and context? That’s a strong indicator of seniority.
The ability to discard the irrelevant and outline one or multiple key focus areas is essential. You’ll never have enough resources to deliver on everything, so you have to prioritize.
Additionally, you’ll never have enough data to ensure complete confidence. So you’ll have to make assumptions.
If a candidate is comfortable with both, it’s a really solid predictor for strategy skills.
Lack of such a skill set will put a tax on the team and the manager, who will have to handhold a PM.
User empathy
Most of the PMs I’ve interviewed never look deeply at the user segments and personas. In certain instances, they even assume “themselves” to be the core user.
The product manager is a customer’s advocate. But this gets lost so easily in the daily work grind.
If there’s no depth or empathy towards a user, a PM’s value in decision making is diminished.
What’s worse, the chosen problems might be off, and the quality of the solutions will suffer.
Ability to work with data
Can a candidate make data-informed trade-offs or rely on gut-feel?
This doesn’t necessarily have to be precise numbers, but just directional inputs (e.g., impact/reach/size of the leverage).
Can the candidate assign relevant metrics and provide a clear rationale for them?
Storytelling
This is often overlooked but quite critical.
What I’m typically seeing with junior candidates is that they memorize the frameworks by heart.
They can even give you a canned high-level response that would be directionally correct.
But what such a response would be missing is a human narrative. It would be robotic and monotonous, checking all the tick boxes. But it won’t score extra points.
What will is the ability to frame an answer in a human-friendly narrative. Stories are how you share information with other people. It’s how you lead and influence teams without direct supervision.
A product manager is a communication hub between multiple functions. Lack of good storytelling will make the job very complex.
🏗️ The perfect response
Product Sense has evolved from the basic 2013 “Goal/User/Problem/Solution” framework (first published in the “Crack the PM interview” book).
A strong modern-day will require you to go way deeper. You’ll have to cover 7 distinct steps, come up with a strategy, make data-informed trade-offs while not forgetting about a MECE attitude (mutually exclusive/collectively exhaustive).
Let’s go through each step one-by-one.
🗣️ Clarify
Interviewer is purposefully misleading you by giving an ambiguous context.
Before jumping into the case, clarify what exactly she wants from you?
Three key questions to ask here are:
Why we want to build this? (what’s the strategic context here?)
When? (are we tight on deadlines? will we need a simple MVP? or do we have a luxury of thinking broader?)
What goal are we trying to achieve? (marketshare/ revenue growth/ retention/ engagement etc).
🎯 Frame the Context
Mission
Be careful with the mission.
Some companies are strongly mission-driven (e.g. Meta). So you’ll have to write it out on a sticker, attach it to the laptop and start every single case with it.
Others have mission as a branding pro-forma and want you to be more pragmatic and jump directly to the goal.
Do you homework and identify which one’s which.
Don’t get stuck too much in the mission though. It gives good directional constraint, but is too high level to make decisions.
Goal
Some interviewers can give you the goal, but others will ask you to figure it out by yourself.
Pick the goal that makes the most sense based on the current state of the product.
Early-stage products are all about PMF, growth (user acquisition/revenue). Mature products can revolve around engagement, retention or even profitability.
👥 User Segmentation
Most candidates stick with the obvious—buyers, sellers, riders, drivers, etc.
Don’t be like most candidates. Take it a step further.
Here are two simple ways to make your answer stand out.
Generational Breakdown
A simple and close to MECE breakdown I typically use is:
Gen-Z/Generation-Alpha - 10-20+ years. Digitally connected. Very tech savvy. Pragmatic. Strong reliance on influencers. Crave fast dopamine. Can give a lot of views/eyeballs, but have the lowest average check.
Millenials (Middle-aged) - 30-40+ years. Digitally educated. Opened to change. Prefer to solve tasks by themselves instead of relying on somebody else. Have one of the highest average checks.
Boomers (Seniors) - 50+. Low to non-techy. Slow to embrace change. Strong work ethic. Trust and rely on authorities etc.
Google such traits and map out your own generational segments. There’s plenty of available studies and data.
Reasons/motivations
An alternative way to map out users is by the jobs they want your product to solve.
But don’t make it overly compliant with the JTBD framework; keep it more humane.
For instance, in the case of “Build a service within Facebook for someone who wants to learn how to play a musical instrument”, those could be:
Learn to play an instrument from scratch as a hobby
Advance the current instrument skills as a hobby
Take the existing academic skills to the next level
etc
Frequency of use
If you’re working with the context of the existing product, a viable way to segment is to use “frequency” as a criteria:
Power users
Regular users
Occassional users
A good rule of thumb is that 20% of power users generate 80% of the revenue and have the highest LTV.
Segment Selection
Once you map out all the segments, assign t-shirt sizes to each based on the:
Reach (Size of the segment) or Average check of the segment
Impact (How much leverage we could have)
We’re going to focus on the hobbyists, who want to learn how to play an instrument from scratch. The rationale here is that ~80-90% of our existing user network are non-professional musicians, who want to learn instruments just for relaxation/fun.
🕵️♂️ Problem Selection
Once you’ve selected a target segment, you’ll need to brainstorm a list of problems.
The majority of candidates end up with 1 or 2. Again, don’t be like them.
Take it a step further and come up with 5-7.
This will closely mimic a real-life case, where you’ll have to make trade-offs.
How to come up with problems?
Problems can be grouped into “hard” and “soft”.
Hard problems are around access, efficiency (time) or money (price, cost, earnings).
Soft problems are centered around usability, convenience, quality.
It’s really hard to find local music teachers with good credibility (access)
Accessible professional tutors cost a lot (price)
Most teachers only give lessons at their home, commutes take a lot of time (efficiency)
Arranging lessons with music teachers is an offline chore (convenience)
How to select problems?
Three key steps:
Arrange them by the level of pain/impact;
Remove the outliers by looking at the goal/mission constraints;
Pick the top ones.
📊 Solution Prioritization
If you have a deadline/constraint, focus on the simpler MVP.
However, if the interviewer hasn’t suggested anything - take the lead and offer a spectrum of solutions:
A very safe MVP
A slightly more aggressive solution
An experimental moonshot (to showcase your “out-of-the-box thinking”)
Safe MVP - A filter allowing you to search for music tutor profiles / allow users to highlight their pages as “tutors” with additional profile info
Aggressive solution - A standalone marketplace for music tutors / music hobbyists
Moonshot - Musical education ecosystem that combines learning/teaching and music publishing.
How to prioritize solutions?
Plot your three solutions on an Impact/Effort chart.
Use common sense and select a Safe MVP as your key solution, keeping in mind that aggressive bet will be developed sometime later.
📈 Metrics
Metrics will heavily depend on the goal behind the case.
You should already know the basics, like acquisition/activation, engagement, and retention.
Take one additional step and also suggest counter-metrics, guardrails, and speak a little about the sensitivity of the chosen inputs.
Limit the number of metrics to only a select few.
Most candidates tend to go too broad and suggest everything that comes to mind. This is noise. Don’t be like most candidates. Only focus on the signal and the key metrics you need to track it.
📋 Summary
Go through the whole case in the reverse order.
Start with the MVP, highlight the key things about the problem, users and the context behind the case.
You might be asked some follow-up questions, improvise on the spot.
🚀 Advanced tips
Timing
Track the time. Ideally, the entire case should take no more than 25-30 minutes.
If it goes beyond that, you limit your chances of success. The interviewer might have extra questions in mind, which she may have to skip.
The main reason candidates exceed the time limit is a lack of structure.
If you don’t whiteboard your thinking, there’s a high chance of getting lost in the logical process.
Leadership
Lead the interview; that earns you additional points.
If the interviewer constantly has to hand-hold you and guide your thinking, you are unlikely to land the role.
Ask questions, but don’t seek validation from the interviewer.
Avoid asking questions that show a lack of conviction or stakeholder dependency, such as, “Do we have a solution in mind?” or “What do our stakeholders want?”.
Feature/business thinking
You can’t jump beyond “mid-lvl” PM if you solely focus on the “features” in the app or the web.
At higher seniority levels, you’re expected to think holistically about the business, the operations, the processes.
Give hints to the interviewer that you can view the problem from the lens of the business owner.
Don’t go full MBA-ballistic with frameworks; just plant hints here and there.
Prepwork
Read and learn about the company vision/mission, decision-making principles, and try to use the heck out of their products.
You won’t immediately speak their language, but trying to understand their perspective better already scores you additional points.
Again, don’t fixate on those principles and company mission during the interview (many candidates do).
📌 TL;DR
Product sense is by far the No. 1 most popular interview question. My analysis has shown that it’s asked in 58% of cases, making it almost a guarantee you’ll encounter one during your interview process.
This question typically comes in three flavours: design a product X for Y, design a Z feature for an existing X product, or pick an existing product X and make it better.
This question allows interviewers to holistically assess candidates in a very limited period of time (~30 minutes). The skills evaluated are systematic thinking, strategy, user empathy, ability to operate with data, and storytelling.
The expectations for this case have evolved over the past 10 years, and having a classic scripted “goal/user/problem/solution” response won’t score you an offer.
The new way to approach it is a 7-step process (Clarify, Set up the mission/goal, Segment users, Select problems, Prioritize solutions, Suggest Metrics, Do a summary). This will require you to think holistically and make data-informed tradeoffs while keeping the whole answer MECE.
Advanced tips to score higher are: Timing (try to keep it within 25-30 mins), Leadership (lead the case and don’t expect validation from the interviewer), Feature/business thinking (show that you can own the business, not just ship features), and Prepwork (learn about the company mission and principles beforehand).