Notes from my Product Management Interviews Part 4: Problem Solving and Data-driven Thinking
This story is written as part of a seven-part series: ‘Notes from my Product Management Interviews’. Read Part 1 here , Part 2 here , Part 3 here and Part 5 here.
In Part 3, I shared how we can approach Product Thinking/Product Sense problems that one would most definitely face in a PM interview. I hope the post was informative and useful and you’d go back to it for repeat reads!
In this part, we will be looking into the following topic:
Problem Solving and Data-driven Thinking
Now, in my experience, I have seen that problem solving is something that comes quite naturally to most of my peers. While some are great with solving a problem in a limited amount of time, some do a great job at it when given some more time or are allowed to set their own deadlines. Interestingly enough, if you can prove your mettle in either way during an interview, your interviewer will be more than happy to take your candidature forward for the next rounds.
Unfortunately, interviews don’t run for more than an hour or so, and thus you need to be a quick problem-solver so that there’s room for a discussion, as well as course-correction if that’s needed.
Problem-Solving Interviews: The Best Approach
Let’s say you walk in to the interview room. You shake hands with your interviewers, and exchange introductions. They give you a pen and a paper, and throw a problem at you: “How will you monetise WhatsApp?”
What do you do? You start throwing ideas: Charge per message. Charge per file shared or by the amount of storage space used. Subscription plans — and so on.
You throw whatever solutions your mind can think of, at that point of time, in that room. But the issue is, a problem like this clearly highlights if you consider the scope of the problem and then think deeply about all the elements surrounding the problem, before jumping into providing the solutions. The more eager you are to talk about solutions, the more red flags will be installed in the minds of your interviewers! Thus, let me walk you through the right approach to solve problems such as this one.
There are various ways to go about being an expert problem-solver during a Problem-Solving interview. Remember, your ability to be a great problem-solver will also define how great of a Product Strategist you are, as some of the companies will specifically call ‘Product Strategy’ out in their Job Descriptions.
One of the best (if not the best) methods is to apply First Principles Thinking to solve a problem. As I had shared briefly in the previous part, First Principles Thinking is all about breaking down a problem to smaller problems and then finding solutions for each smaller problem, then assimilate all the solutions to come up with a fantastic solution to a problem. Sounds simple, but to think this way, again, it takes practice.
Adding some links that’d help you understand what First Principles Thinking is all about and how it is different than thinking by analogy (which is the more common way of addressing a problem — again something that may not yield great results in a problem solving interview).
First Principles: The Building Blocks of True Knowledge (Farnam Street): https://fs.blog/2018/04/first-principles/
I’d recommend reading this blog post multiple times, till you can talk about the theory behind this method confidently, say, in a social gathering. After that, I’d recommend making your own quick notes that could help you remember what First Principles is all about. It will give you something to fall back on, every time you have a problem solving interview to go to, and will also help you build 100% trust that this method will always be effective if done right.
Next, this video will help you visualise the concepts around First Principles thinking. Be patient when you watch it, a fun example comes up somewhere around the 4th minute: https://www.youtube.com/watch?v=HZRDUZuIKg4
Once you have understood how to apply First Principles thinking in problem solving, I’d recommend you to watch the following video on YouTube, that takes a bigger problem and breaks it down to smaller problems (the essence of First Principles, basically), and represents the same in a fishbone model. Enjoy the video but remember not to pay too much attention to the naming conventions of the framework used. Nobody will score you higher in an interview if you can name a framework correctly; your skills around how you apply a framework is what will help you make your responses extremely structured and thus, take you places!
How to Solve a Problem in Four Steps: https://www.youtube.com/watch?v=QOjTJAFyNrU&list=PLdECGQUWSsoaqmiokbDUB5F3XAz9je6Vf
And so, keeping the First Principles thinking approach in mind, let’s go back to our previous problem statement — “How will you monetise WhatsApp?”.
The best approach would be to ponder around the questions such as: Who is the user and what do they need? Why would a user pay for a service like WhatsApp? For the value, right? What value does WhatsApp provide to a user? How does it change when it becomes easier to use?
Initial questions like these will help you start off on the right track to begin solving the problem in the best possible way. At this point, you can ask your interviewer for feedback and see if they agree with your line of thought.
These initial questions will help you understand what really could be reasons that is not leading to WhatsApp not monetise at present. It doesn’t matter how brilliant your solutions are — as long as you are asking the right questions, you are showing off your thinking skills to be driven by logic and creativity. Thus, your interviewer would recognise that you’d like to spend a significant amount of time understanding the problem statement and the factors that contribute to it.
The following steps finish off the problem solving process for this particular problem statement:
Step 1: Flip the problem statement- WhatsApp is not currently monetising.
Step 2: Ask the question: “Why is that the case?”
Then, break down the problem by coming up with the possible sub-problems such as these:
a) do they want to focus on gaining and retaining their customer base by keeping the app free?
b) are they keeping it free for regulatory reasons? and
c) since their main feature is to exchange messages through chats, do they not want the users to leave that feature?
Step 3: For each of the sub-problems come up with a possible solution:
For a) a solution could be: tie up with business partners that’d provide more value to WhatsApp users, integrate them to WhatsApp so that WhatsApp users can avail their services through WhatsApp, and then charge these partners based on every successful transaction. Example of such partners could be: food delivery service providers, cab-booking service-providers, house-hunting service-providers, etc
For b) a solution could be: work with the Government to address regulatory challenges such as data security and privacy challenges, then safeguard PII information and, then, by using only behavioural information, serve ads to users — charge businesses for ad-spaces and other ad metrics(eg- no. of ad-clicks, no. of ad-views etc) on WhatsApp(similar to Facebook’s ad model)
For c) a solution could be: create chatbots to sell products which need higher degree of conversation and persuasion, such as insurance products, and then charge the businesses whose products will be sold through chatbots
Step 4: Summarise by discussing what you have highlighted in the previous steps and ultimately wrap up with the solutions that you have recommended in Step 3.
Data-driven Thinking
Data is everywhere around us, and most of our day-to-day decisions are directly or indirectly influenced by a corpus of data that we would have at our disposal. The problem, as you’d have read in many posts on the internet, is that we might end up using data that would simply lead us to a wrong path, and thus, lead us to draw a wrong conclusion. There are various reasons that might cause this:
- too much of skewed data everywhere: This post explains its effects and how to handle it: https://towardsdatascience.com/skewed-data-a-problem-to-your-statistical-model-9a6b5bb74e37
- too much of uncleaned data
- un-categorized or unclassified data
….and so on.
I asked my friend Ram Choudhary, who is a Product Manager at WorkIndia, to share his thoughts on data-driven thinking. Adding an excerpt of his response below:
“Let’s say you are a PM in a fast-paced startup-like environment. Your business goal is to increase the average revenue per user by 10% over the next 1 quarter. Where do you begin? Do you just throw something at the wall and see what sticks? You could try that, but it’d be clear to everyone that you are making wild guesses with your experiments..
So, wouldn’t it be better if you got hints? Looking at data is that hint. Being able to read the data in the right context, is being able to understand the hint right. For example, I query the databases frequently to analyse and understand what’s happening on the app, how things are comparing, is there a trend/pattern I am seeing, etc.
When you are able to do this on your own, you can answer your questions that can help you solve the problems better. In this way, data helps you in problem-solving. Over time, with some hands-on practice, you’d be able to ask your BAs for great analysis. They’ll do a phenomenal job of putting in all the hard work to find the data and present it in the way that you’d need, but you need to know what to ask first.”
Thus, as a Product Manager, while you won’t be expected to be a proficient data scientist, you will be definitely expected to do the following:
- identify what data is right for you for a given problem that you are trying to solve
- understand the language of a data scientist/data analyst in your team (remember, Stakeholder Management!) when they present you with interesting data as part of a project that concerns you
- draw insights from given data or data visualisations: I cannot stress on this enough, as I have seen in my past experience as a Product Manager, the better insights you derive (by connecting the dots in the data you have at your disposal — using creativity), the better you can define a path that helps you successfully take your project to completion. This medium post talks in detail about Data in Product Management where you can read about deriving insights, product data, and much more. As always, don’t get caught up too much on terminology, just deriving the essence is what matters, especially during an interview: https://towardsdatascience.com/becoming-a-data-driven-product-manager-f320d0182a9c
Finally, during your Problem-solving interviews, you will be needed to talk extensively about deriving product/feature data and looking into the same when you are addressing a large problem. Pro-tip: wherever your interviewer asks you the hows behind a solution you are proposing, almost always take it as an indication that they want to see if you will be relying on data to make that conclusion.
In addition, understand that data doesn’t always have to be highly technical, what is generated on telemetry or on system logs, data (especially during interviews) could also be broad-level content or response that you might obtain by simply talking to people and asking them specific questions (for example: market surveys). Like I mentioned above, data is everywhere around us!
I am adding some of the Problem Solving questions that I have faced in my PM interviews till date. In most cases, the previous interviewer or the HR had informed me before the interview, that it will be a Problem Solving interview. In any case, don’t get lost too much in thoughts with regards to how an interview is being named — just do your job in being a great thinker and a fine problem solver, and the rest will fall in place! One more thing, sometimes these problem solving questions could be Product Design/Improvement questions in disguise — the more you practice, the better you will be able to identify what a problem could be all about.
List of Problem Solving Questions
- Instagram is seeing a steep drop of users over the past two months. As a Business head, what do you think could be the possible reason(s)?
- Suggest improvements in Instagram. How will you prioritise them?
- Imagine that you are a Product Manager at Uber during its initial days. How will you decide on the pricing of the cabs that needs to be shown to a user who searches for a cab on the Uber app?
- Imagine your company makes bulbs which are durable enough to last 30 years. How will you price these bulbs (mention the final Price and assume that these bulbs will be sold in India)?
- How will you monetise WhatsApp? (we have solved this above)
- Imagine you work in Ola as a Product Manager in its early days. How will you introduce Ola Auto in the market?
- Imagine you are a PM at Uber. You notice that drivers are cancelling rides after accepting them. How will you solve this problem? Highlight your approach and steps.
- You are a PM for Gmail. You see that there’s a significant drop of users over the past few months. You hypothesise that the users are moving to Slack. How will you test this hypothesis? Highlight your approach and steps.
- Flipkart want to get into a category, say, Pharmacy. Imagine you are a Business Manager/Product Manager assigned to explore the same. How will you go about validating 1. whether or not we should introduce this category, and why? , and 2. if yes, what should be your approach and the steps involved?
So, that’s it for this topic! I hope you liked reading this post, and will find this post useful to prepare for the Problem Solving interviews you face.
Feel free to reach me at contactsoumyam@gmail.com for feedback, questions or ideas. If you liked this post, do add claps below — it is a nice ‘vanity metric’ for me that I am doing things right with this series. ;)
Thanks for reading!
Disclaimer: credits to all images used in this post are attributed to the original source in the captions. The author takes no credit for the awesomeness of these images and consequently all copyright claims should stay within the original source’s boundaries
This post has been published on www.productschool.com communities.
