Product Manager Portfolio
Back to questions

Diagnose a 15% drop in Instagram engagement.

Additional information provided after inquiry from the interviewee:

  • This drop in engagement happened drastically over the course of 2 weeks
  • No specific location was given
  • The metric used to measure engagement was the number of posts created, posts does not include stories or reels.
  • Related to Problem Solving
  • For Consumer PM
  Meta
Add Your Answer
Answers (1)

Context

 

Before we dive into diagnosing this, let’s just recap what Instagram is and why users use it. Instagram is a UGC platform where users can share and view photos and short form videos from other users including their friends, celebrities, and brands. Users enjoy sharing what’s going on in their world as well as seeing what’s going on with everyone else which is what Instagram is seeking to accomplish as their mission statement is to capture and share life’s moments. A drop in engagement means less of life’s moments are being captured and shared. 🙁

Clarification

I think it’s worth spending a moment to make sure we understand the question properly.

  • It may go without saying but let’s confirm that we’re confident in the accuracy of our metrics. Nothing feels worse than finding out the data you’ve been using to guide your decision making is invalid to a certain degree.

  • There are natural fluctuations in user behavior and the metrics that measure that behavior. Let’s confirm that this drop is statistically significant and not just attributed to variance.

  • “15% drop in Instagram engagement” is actually a fairly ambiguous statement. What are the exact metrics or set of metrics we are using to observe this 15% drop? Is this drop in engagement amongst viewers, creators, or advertisrs? (Interviewer: 15% drop in the number of posts created)

Honing in on the issue
Great, let’s try to diagnose this further and really hone in on the issue.

  • When did we notice this drop in engagement? Was this a sudden sharp drop or something that occurred gradually over time? How long has this drop been going on for?

  • Is this drop in engagement uniform across geographies or are their certain cities or countries where it is more or less observable?

  • Are there any user segments or use cases where this drop is more prominent? For example, maybe users are engaging with the feed less and spending more time in Reels.

  • Is this drop uniform across browsers and devices?

Interviewer: This drop happened suddenly over the course of two weeks.
Potential Causes
Now that we have a better understanding of what exactly we are seeing amongst our users, let’s brainstorm some potential causes for why it might be happening and how we would test it.

  • There could’ve been a 15% drop in Instagram usage overall which would lead to a 15% drop in total number of posts being created as well. -> We could look at the total number of users on the IG platform.

  • Product changes could’ve de-incentivized users to create posts. For example, maybe Reels is featured more prominently now so creators make more Reels instead of posts and feedback on the posts they do make has gone down as well. -> Are there recent product releases? Can we examine the data from their a/b tests to see if engagement also dropped amongst users who did not prominently experience Reels.

  • Has anything happened to draw creators away from our platform? For example, if TikTok increased the payout of their creator fund to creators that would be an issue for us. -> Market research for recent news from competitors and see if it maps to our timeline.

  • If it was geography specific, it could either be an infrastructure issue like a lot of Texas lost power during a cold snap last year or a censorship issue. Maybe Russian banned creators from IG as a pre-emptive move against propaganda during their invasion of Ukraine.

Summary
Through our investigation we established that there over a two week period there was a 15% drop in the number of posted created on Instagram. We proposed several hypotheses for why this might have happened and how to validate or invalidate them including: decrease in overall users, product cannabalization, increased competition, region specific regulatory or infrastructure issues.