How to diagnose a problem as a Product Manager

If you are a Product Manager, then you can easily relate to this situation — you are getting ready to watch a movie on a Saturday night and suddenly there is a Slack notification. You open it and see there is a drop in a very important metric and everyone is suddenly online and you have the responsibility of diagnosing the issue. It’s very crucial to solve it and send a hotfix, if required, as soon as possible. How would you approach the problem?

Okay! Let’s take an instance.

Suppose you are a Product Manager at Snapchat, and the problem is 20% engagement drop on the Snapchat

If I was a Product Manager responsible for diagnosing such an issue, I’d ask questions like:

  1. How do we define engagement?
  2. Is the drop gradual or sudden?
  3. Is there a particular age group that is more affected?
  4. Is there certain geography that’s affected the most?
  5. Is there something else that was launched that might be drawing engagement away from the App?
  6. Did something change recently in the industry that might lead users to drop?
  7. Is the drop noticeably larger on a certain platform, for example, iOS vs Android?
  8. Were there other features affected?
  9. What percentage of users are reporting engagement drop?
  10. Were there fundamental changes in how we capture engagement since refactors?
  11. How is the hourly trend on a daily basis looks like?
  12. Is the drop on a particular version?
  13. Was there any related backend deployment made?
  14. Was there any edge case that is now more prominent?

Also for each of the above questions, I’d ask manyWhy wherever applicable to get to the root.

Also, do not assume anything. Question every assumption.

So now you have an idea. You have to bombard a series of questions before even looking for a solution. Ask as many relevant questions you can ask.

A first principle is a basic proposition or assumption that cannot be deduced from any other proposition or assumption

These will help you get to the root cause quite quickly and you’ll be certain of the solution you get.

If you have 6 hours to find diagnose a problem, spend 4 on the question.

Product Manager | Stock Market Trader. I write about product management, financial freedom, and personal growth.