Facebook's DAU is decreasing 15% per week from the last 3 weeks. Find out why.

Clarification of the metric on how DAU is calculated? Should I assume as users coming on the FB platform everyday and spending a specfic amount of time, say 45 seconds OR taking 1 engagement action like post/like/comment/share OR both?

 

First thing I’d like to get ask some more clarification questions that can be easily found by our internal data. Once we have that then I will come up with some hypothesis, how to test those in order to diagnose the problem. How does that sound?

 

1. Look at is whether it is sudden or progressive -- It looks like it is progressive as it has been happening for over 3 weeks.
   1. Is it seasonal? Say around holiday time when people are mostly spending time outdoors? -- Assume first time.
2. Next I want to check whether there was any change in the query logic feeding into the metric or change in backend tables/infrastructure that is not logging the data correctly to rule out any data quality issues
3. Check whether there is any specific market the drop is coming from or is it globally? -- Say globally
4. Check whether any specific age group? -- Say young adults (age 18-29)
5. Any specific platform (mobile/desktop) or both are equal? -- Say both equal
   1. Any specific OS (android/iOS) or both drops equal? -- Say both equal
6. Decompose the metric to see where the drop is coming from:
   1. DAU = Existing + New - Churned
      1. Say existing
7. Internal product changes:
   1. Made any significant changes because of which we are noticing Primacy affect, that is, change aversion and users are resisting that change? -- No
   2. Doing any sort of AB testing that is causing people friction to come back to platform? -- No
   3. Any bad PR about Facebook, say accounts compromised or any legal issues because of which people are boycotting FB -- No
   4. Cannibalizing: Spike in DAU in last 3 weeks on Insta/Whatsapp? --Noticed spike in Insta
8. External:
   1. Hard to get this data but is this a social media trend going on where people trying to be more active outside social media -- Don’t know
   2. Any new/existing product in the market where people are spending more time? -- Don’t have external data
      1. Cohort by age group. Say young teens are spending more time on Snapchat or young adults are spending more time on LinkedIn.
9. Hypothesis & confirming:
   1. Young adults have started using more Instagram. This can be found out by looking at the DAU, avg time spent on Instagram.

Clarifying Questions:

1. What is meant by active? Is it users who log in and engage in passive viewing or users who are logged in and perform an action or set of action/s?

2. Is the metric drop seen for a specific cohort or a particular demography?

3. Is the drop observed in a specific device (Mobile, Desktop, Tablet, Laptop) or on a particular platform/s (Web, iOS, Android)

4. I'd also check the DAU/MAU a.k.a stickiness to check if our MAU's have been affected?

5. Have we seen a similar incident before? Maybe its a trend? Ex. People are generally busy with festivities during Thanksgiving and might not be socially active?6. Is this drop due to a major event that happened during that time? For example, there was ongoing controversy with Facebook, people boycotting social media in general or a major calamity 

For this case, I'm assuming that the drop is across the USA and for Android users. MAU (in totality) hasn't seen a significant drop. The definition of the active user: a user who has an active session, a user might not engage in comments/posts or likes. Even if the user is surfing and viewing the content. It will be considered an active user. Here is the list of issues I would look into:

1. Did we roll out a new version on Android that is causing the user issues while logging in? Or is this due to our application working differently with different OS's? We could delve deeper and distinguish the user experience among them.

2. Did we roll out a new feature that users are not liking, therefore, choosing not to use Facebook?3. Was an existing feature modified, for example, a revamped UI/UX layout that users aren't liking that much?4. Check if this drop is due to an ongoing experiment with the Android users (probably A/B testing) and check with the team who is conducting it5. Check with the Engineering team to see if there is a technical glitch in the tracking metrics measuring user engagement. I'd also check if there was a latency occurring due to new features which are specifically not working well with Android6. Did we start showing too many ads on the platform causing users inconvenience? We can check this by the Ads/Newsfeed items ratio per DAU and per MAU to notice the difference

7. Do we have a competitor in the US market which the Android users are preferring over ours and carry out an analysis accordingly?