I wrote earlier about the Google PM Interview Process & Preparation. That article, although lengthy, was focused on the overall interview and recruiting process. This article is going to be much more targeted on the preparation aspect for the interviews.
Although I focus on the Google PM interview, I hope this is also helpful for folks considering interviewing for PM roles in other tech companies, and even for those interviewing for other related functions.
You will get a mix of questions throughout your PM interviews. These will range across many different areas, but you can certainly expect a few of the following types of questions:
- Strategy (“should Google compete in market [X]?”)
- Design (“design an elevator for blind people”)
- Analytics (“how much revenue does company [X] make”?)
- Technical (“how would you design an algorithm to detect duplicate calendar invites.”)
The variances in questions are huge. Your temptation will be to try to prepare by answering a million different questions and getting to the right answer. This is counterproductive.
The real challenge will be structuring your answers
In my opinion, the secret to succeeding in the interviews lies outside of the realm of whether your answer is correct or not. Instead, you should focus on delivery of your answer.
What is your interviewer trying to assess? What is your interviewer likely NOT interested in assessing? Here some things I think are most important, but note that each interviewer and role may be slightly different. However, IMHO, these are directionally correct.
|Interested in assessing
||Likely NOT a focus
|What’s your thought process when breaking down a fuzzy problem?
||Do you get to the ‘right’ answer to my question?
|How do you approach a very challenging design question?
||Do you come up with the best ideas right in the spot, within a 20 minutes interview?
|What’s your collaboration approach when working with engineering?
||Can you code? (typically)
|How do you balance long term vision with short term needs?
||Can you be the next Google CEO?
|Do you understand technical challenges that a proposed solution may have?
||Can you design the best technical implementation for a problem?
Given that, focus your preparation on how you will communicate and how you will structure your answers.
The worst thing that you can do is to give an answer to an ambiguous question without walking through your thought process on how you arrived at the answer.
The interview is not about getting the right answer, the interview is an opportunity for you to show how you get to that answer and your thought process.
Build your own framework
There are plenty of frameworks to structure your answer. Often cited is the CIRCLES method for design questions. I’d suggest that you read through a few frameworks and build your own. Here are the core items that I think are crucial for your success.
Understand the real problem: the question you are asked will likely be ambiguous. Don’t try to answer right away. Think about ways in which you can clarify and make sure you understand the question in detail. Also, understand the goal. What are you trying to accomplish? What are the constraints that you have? What should you consider as ‘in scope’ or within the realm of possibilities for your solution?
Question: “How would you improve Google Search?”
Are we trying to improve revenue? UX? Increase daily users? Repeating visitors? Are we focused on the Desktop version of Google.com? Mobile? Google Home assistant? Is it in ranking the top results? Or on giving answers within the page itself? Or are we focused on improving the ads experience? How much budget do we have for the solution? Is there a particular cost revenue/cost target?
There are a gazillion questions you can ask to narrow down the problem. I’d ask a few of the more crucial ones that will allow you to focus on what matters and understand the goal. If your interviewer readily engages with you in providing additional information, great, keep going. If not, take that as a sign that they expect you to make reasonable assumptions and explain them as you work through the problem.
Structure your solution: Think of how you will solve the problem. Sketch out your solutioning approach. Think of the core elements of how you’ll approach it. If this is a product question, map out the user journey and identify key pain points. Identify your various user personas and how you will prioritize one vs. the other. For technical questions, think of the various alternatives you can think of. Think of complexity of your implementation — what data is needed? Cost of keeping up the feature? Think of privacy concerns.
At this point, you want to set the overall way you will tackle the problem, you are not yet answering the question or solving the problem.
At this stage, you want to present how you will tackle the problem to your interviewer and get feedback. For example, you could say:
To figure out how I’d improve Google Search, I’ll first pick a user segment that I think it’s crucial. I’ll then map out their objectives and key pain points when accomplishing that goal. I’ll then propose a solution that will tackle the key pain points and discuss some alternatives I can think of. Once we agree on a solution, I am happy to discuss the implementation approach I will take.
First, however, I want to clearly establish the goal we are going after, which is X. Knowing that X is our goal will help us narrow down to the right user segments and problem types to solve.
I’d strongly encourage you to write down your framework on a large piece of blank/white paper as you’re talking through it and write down the core pieces of your solution (pain points, user personas, core constraints, etc.) and refer back to it as you work through the solution.
Your goal at this stage is to be structured. Remember, you’ll have 10-20 minutes to walk through your solution and you don’t want to be all over the place. A structure is crucial for this.
Propose an option and an alternative: here is where you work through your framework, as you’ve explained it in the last step. You’ll actually identify a persona, identify use cases and user flows, identify unmet pain points, etc.
The core things that I’d stress about this stage is that it’s crucial that you communicate clearly.
- Don’t just jump through to a solution. Instead, talk about how you’ve arrived at that solution.
- Don’t simply say “I’m going to assume X”. Why do you think X is a good assumption. How would you validate or double check in the future if X is, indeed, a good assumption. If there isn’t a way you can think of to validate X, how will you test for it or minimize the risk of getting it wrong?
- Write down key reasons why a particular part of your solution may be difficult to accomplish. Discuss (and write down!) ideas about alternatives you may consider to mitigate some of those difficulties.
Evaluate the pros/cons of your solution. Consider the components of an ROI here. What are the costs, complexities, and risks of your solution? What are the potential benefits? How do you get the most of the latter while minimizing the former?
Talk about implementation & potential complexities and challenges. Time permitting, discuss how you’d implement a solution. If this is a technical interview, pseudocode may be great at this stage, or a general framework of the algorithm component, data sources, etc. that you’d tap into. If a design question, consider sketching out a solution / design or writing down Critical User Journeys and key requirements. Call out complexities of the solution. If it is a strategic question, get tactical about how you’ll implement it. How will you go to market? What parts of your strategy would you implement in the short term, medium term, long term (draw out a roadmap!). How will you mitigate key risks in personnel, execution complexity, costs, PR, etc?
Finally, wrap it all up by discussing next steps. Refer back to your pain points. Discuss, how have you solved the pain points? How have you solved the problem? Moving forward, how will you measure and make sure that your solution works? What will you measure to determine whether you should continue to go down the path you’ve outlined or you need to pivot? What signs will you get that your solution or proposal is good? How will you validate key risks you outlined previously.
How to practice
The above was a very long way to explain what you should be focused on when preparing for your interviews. I focused on the above since I don’t believe it is productive for you to focus on fundamentally acquiring new skills or knowledge you are completely unfamiliar with to pass the interviews. The interview questions are complex and varied enough that such preparation will likely yield OK but uninspiring answers. Instead, prior to an interview focus on what may actually get you from OK answers to an excellent discussion with your interviewer. Focus on what can get you to shine.
Practice with a friend. The best way to practice is to have a friend who has gone through the process practice with you. Ideally another product manager, since they’ll have a ton more context on what makes for a good answer.
Practice by recording yourself. In addition to practicing with a friend, record yourself using your phone/computer as you answer practice questions. Then play back the recording. This is way more helpful than people initially expect.
Review fundamentals. While I think that you’ll succeed mostly by capitalizing on your existing strengths, as you get stuck answering a sample question review the fundamentals behind it. If it’s a technical question, review the algorithm, data structure or infrastructure concepts that would have helped you. If it’s a design question, run through a few design frameworks that may have helped and tweak your overall response framework.
Get plenty of sleep! Personally, I perform 10x better at any task when I’ve gotten plenty of rest the two nights before. I suggest that you prioritize sleep over obsessively reviewing things in the last stretch before your interviews.
Don’t forget the soft stuff
I didn’t talk at all in this article about the soft questions, such as “why do you want to work in PM?” or “why Google?”. I also didn’t touch upon behavioral discussions such as “tell me about a time in which you faced difficulty, etc.” Be prepared to answer any of these, but don’t stress too much about it.
The only tip I’d give is to pick ahead of time a few projects or efforts that you’ve worked on in the past that show your diligence, ability to communicate cross-functionally, rise up when there’s a challenge, and collaborate well. Review your past accomplishments and challenges so that, if a question like that comes up, you have fresh in your mind a list of potential activities to talk about.
I’ve seen incredibly talented people not get a job offer after going through the Google PM interview process. The process is optimized to minimize false positives, that is, sometimes even if a candidate seems ‘good’ they may not get an offer if the hiring committee has doubts. Some of these folks that don’t get an offer in the first try interview again, usually 1+ years later, and do get an offer, so make sure you keep that option in mind.
Do I need to code during the technical interview?
While there’s no hard rule I have read about online, for PMs the focus should be in demonstrating:
- That you recognize technical challenges and scalable vs. non-scalable solutions
- Can smartly talk through how you’d design an algorithm/solution to a problem
- Understand data structure and internet infrastructure fundamentals
I haven’t heard about PMs being asked to code during interviews, although pseudocode does seem to come up.
Cracking the PM Interview*: best ~$15 I spent when preparing.
Thepminterview.com: specifically their Estimations, Product Design and Strategy questions.
My article on Google interview process: for a general overview of the process.
Update: this article is now also available as a Google Doc (link) so you can easily add comments and help us improve it.
* Note: I use some affiliate links in this post, marked with an asterisk. If you click through and purchase products I earn a small referral fee at no cost to you.