Cracking your next PM Interview: Problem Solving Round

If I were given an hour in which to do a problem upon which my life depended, I would spend 40 minutes studying it, 15 minutes reviewing it and 5 minutes solving it — Einstein

While most companies will name this as problem solving round, this really is “approach to problem solving” round. Nobody is expecting you to solve the problem in that 30 min time-frame. Instead they are interested in the approach you are taking to get to the root of the problem, how “out-of-the-box” you can think on solutions and finally how will you prioritize the solutions you have listed. All 3 are important elements, solving the problem in itself is of least importance.

Have a 5000 feet view to solve the problem quickly

Types of problems you can expect: Most of the times questions are around a certain metric movement or for a specific problem. For eg.,

“Google Maps had a 10% drop in daily active users. What do you do?”

“There’s a 50% drop in the number of Uber rides per day. Diagnose the problem.”

“Reducing cancellations by Uber drivers”

First 30 secs: As soon as the interviewer has finished asking the question, 8 out of 10 people will start blurting out the solution or assumptions(at least the ones I have interviewed). If you are one of those 8, save yourself and interviewer some time by ending the interview right there.

Spend a lot of time in the problem space rather than in solution space. This is what you are doing as a PM in your everyday work. And this is what people are looking for. This is what your mind should immediately start thinking…

1. Is this even a problem? or how do we know this is a problem?
2. What is the scale of the problem? Is the problem worth solving?
3. What are my assumptions in trying to understand this problem?
4. What will customer get if I solve this problem?

If it comforts, most of the times even the companies don’t know how to solve the problem — so again please don’t focus on solutions. Practice these problems from the Product Management Exercises

Step 1: Know the problem — 5 Mins

1. Clarify the product and metric: Although you would know the product, for eg., Google Maps, its a good idea to mention your assumption. Similarly clarify the definition of metric as well.
2. Clarify the goal: Why do you want to fix this problem? For eg., Fall of DAU is resulting in reduced ads revenues. So fixing this problem will help the company plug this revenue gap.
3. Clarify the source of the problem: How did we get to know that there is a 10% fall in DAU? Is the source verified to be working fine?

Step 2: Breakdown the problem — 10 mins

1. Extrinsic factors:Many times people assume if a metric is not doing well, it is an indication of a fault in the product. This is not true. Even if your product has been working fine, external market forces can cause the rise/fall in metric. This true when some metric is doing extremely good too. So you must make sure that there are no external factors at play here…
   — Time duration when the metric was measured? For eg., in case of our Google maps problem, is the reported problem around christmas holidays where users are not travelling to office? eliminate seasonality
   — Impact of a particular geography? For eg., was there a fall in the DAU because of turmoil in any specific country
   — Any outage that was reported?
   — Any new entrant or substitutes in the market? “Did Facebook introduce a maps product?”
   — Any policy changes? “Did iOS introduce any policy change that scares users away from sharing location?”
   — Any impact of partner eco-system? “Was there a strike by Uber drivers?”
   — Is the problem unique to our product or in general industry is facing this? “For eg., have the number of people driving cars reduced because the fuel prices have gone up significantly?”
2. Intrinsic factors: If you have concluded that problem indeed seems to be with the product, proceed with understanding the intrinsic factors impacting the metric…
   — Did the source of traffic mix change? App/Web — iOS/Android — paid vs organic
   — Did we see the change in metric in any particular platform? iOS vs Android, Chrome vs Mozilla, MacOS vs Windows
   — Did we see the change in metric in any particular customer segments? Office goers vs vacation planners, New vs Existing users, Male vs Female customers, teenagers vs 40+s
   — Is there a counter metric we are looking at? “If DAU has gone down, did we see a rise in crash metrics?”
   — Any product upgrade that is causing the problem?” The “go” button went missing”

This is a note I wrote on metric modelling

Step 3: Digging deep — 10 mins

Once you have zeroed in on the likely causes of the problem, connect the dots and relay your findings so far to the interviewer. For each of the zeroed in likely causes try to either call out your assumptions on what might have lead to the cause or ask the interviewer explicitly on details. You should be running the 5 whys in your head now.

1. Build the funnel view: Now that you have the likely causes and the assumptions behind it, try to walk through the entire customer journey
   “When customers such as ____________ (customer segment) are trying to fire up our Product in _______ (device/browser), they are don’t have or are missing ______ (features), this is contributing to downfall of our DAUs, ultimately affecting our revenues”
2. Bring the attention to input metric: Now that we know what is causing the problem, identify an input metric that, when improved, will eventually impact the outcome. Position this input metric as a success metric for the solutions you are listing out. For eg., if the problem given was to reduce the fake news spreading on twitter, one possible input metric will be to increase the verified accounts on twitter.

Now you have distilled a very vague observation to actionable, result oriented problem statement.

Step 4: Listing Solutions — 5 mins

Now that you have an agreement with the interviewer on connection of input metric to outcome, you should be laser sharp focused on improving the input metrics.

1. Go broad: In design thinking parlance this is the stage where you should diverge and list out broad set of ideas.
2. Constraints: For each of the ideas be mindful of what could be some of the constraints
3. Strategic vs Tactical: Differentiate the ideas on whether they are strategic in nature ( fixing for long term) or tactical ( immediate actions that will stop the bleeding)
4. Pros n Cons: For each of the ideas also keep track of pros and cons that are possible.

Pro-Tip: The more you keep practicing the problem solving questions, your ability to think quickly and generate more ideas keeps improving.

Step 5: Prioritising solutions — 5 mins

If you have reached till this stage smoothly with interviewers attention still intact, this is that stage where “Mario now jumps and pulls the flag down”. Unless you make a blunder, most of the times interviewer may want to understand how you are prioritising your ideas.

Generally followed approaches of MoSCow, RICE or any other method works okay, but the best way is to put the idoueas on a 2X2 matrix of impact and effort. Important tips…
— Don’t grossly miscalculate the impact of your ideas or the effort required to execute
— Don’t argue with the interviewer when alternate views are presented
— Don’t be submissive, its okay to stand on your views — but politely
— Be open to new use cases opened up at this point. Can happen in real life too
— See what differentiates you from competition, anything that can give you competitive edge

Step 6: Summarise — 1 min

Make sure that you relay the entire conversation with problem statement, cause/correlation, input metric, solutions, prioritisation and final recommendation.