Daily active users on Instagram have dropped by 20%. What are the steps you will follow to identify the root cause?
You'll get access to over 3,000 product manager interview questions and answers
Recommended by over 100k members
My thoughts on this Facebook problem solving question.
Interesting. Okay, so daily active users of Instagram declined by 20%, which seems super significant. My goal is to identify why it happened so that we can fix it.
My proposed approach is to first rule out obvious reasons and then define and break down the metric so that we can isolate the issue.
If I were a PM on Instagram, I would surely be aware of seasonality patterns such as (1) weekends being the highest usage days and (2) holidays being highest usage seasons. So I assume the reason behind the drop is not seasonality related. Second, I would assume that if there as a broad service outage the whole company would already be aware of it and working on fixing it. So I also assume this is not the case. Does that make sense?
I assume that by Daily Active Users you are referring to Users who open the Instagram app at least once on a given day. Is this a fair definition? Cool!
Next, I would start breaking down Daily Active Users by a series of variables. Specifically, I would break it down by:
- User type (this will help isolate whether the issue is with e.g. with the registration flow):
- New
- Existing
- Traffic Source (this can pinpoint whether there is an issue with any source of traffic e.g. notifications system malfunction):
- Direct traffic
- Notifications traffic
- Search Engine traffic, which shoud be further broken down by SEO and SEM traffic
- Referral traffic (e.g. from Facebook)
- Other
- Platform (was there any new feature deployed in any platform that could have caused the 20% drop?)
- Mobile Apps:
- iOS
- Android
- Web
- Windows
- MacOS
- iOS
- Android
- Mobile Apps:
- Region (to make sure the issue is not with a specific geography)
Clarifying questions:
Assuming it is an app based issue currently..
Was this 20% drop a drastic fall or has it occurred over few months?
Is a similar trend observed in Wapp, FB as well?
If they are not daily users anymore, have they completely abandoned the app? If not how often are they using the app and which feature?
Is there any new feature that has been introduced/old feature that has been discontinued,post which this drop has been observed?
Out of Instagram feed,stories,ads,IGTV features, which were the ones being used by these users previously before they went inactive?
Are these users of any particular demographics/observed in any particular geo?
Probable High Level Reasons:
- User Engagement has become low which is causing a set of people to go inactive
- The users are of a particular age group who are not finding anything relevant to share with their connections
- The daily active users who used to share travel photos and videos are now unable to do so because of external factors like Covid scenario
- A feature introduced into instagram is not being accepted well by 20% of such daily active users
- Users are switching on to a competitive product which has newer features and people are using such hashtags in their wapp status/fb stories
Now, I would prioritise the reasons which are internal to Insta and work on those first.
Hypothesis
1. The DAU number is impacted by lower engagement among users of a particular age group/interest area
Metrics: #followers of the DAU vis-a-vis reaction received on their feed/stories, #story reaction/replies vs #story viewsfor the users who dropped off,#profiles followed as per suggested profiles,# times their profile were suggested on others' insta account,#times they used the most used feature on insta,#interactions with feed/stories of profiles being suggested by insta
2. The DAU number is impacted by a new feature being introduced on the app which is either not easy to use or not received well by the audience
Metrics: #of times the new feature has been launched by the user vs #of times it has been used, #of interactions with followers on the output of this new feature when compared to other feature based posts, % change in follower count before and after this feature use
These metrics would help me reach the Root cause of this issue.
To conclude, if the drop has been drastic- it mostly points toward a sudden change in a feature/usability. This might be feature addition/roll back or a change in the algorithm causing a change in the performance of the app. The other reason can be entirely external/unprecedented.
If the change has occured over time, for say a few months, it means- the users have gradually lost interest in the app because of lesser interactions with connections/irrelevant follow suggestions/no increase in followers/drop in content shared by connections/not being able to share new content with followers.
Q: DAU is users opening the app or is this mainly users are opening the app but not posting anything? (not posting anything. It doesn't do much good if people open but don't post).
Me: Well, I agree and i don't at the same time. People opening the app and not posting doesn't mean people aren't scrolling through feeds, engaging with it, etc. and that results in business impact. But yes, over time, if new posts aren't generated then it will impact people scrolling through and we will loose even retained users.
Q: Has the dropped happened over a course of tiem or suddenly overnight? (since about a year). That's a great info to keep in the bank for now.
Q: Is the drop noticed on any specific device types such as OS specific? (no)
Q: Is the drop worldwide or just US? I ask this since it could lead us to think of seasonality, holidays, natural calamity like Katrina (but not harsh winter in Austin), major power outgages, etc. (worldwide and yes we could eliminate other aspects).
Q: So you said people are opening Instagram but not posting. (correct).
Q: Alright, what I would like to do here is explore some areas with you: 1)Any code changes 2)UI changes 3)News that could be impacting us 4)worldly events. Some others may come to mind as we proceed.
Interviewer: Great, let's talk through it all. What questions do you have?
Me: well, right off the bat I do have another cursory question to understand the problem. Is there specific type of posts we are thinking of here? Like we don't have as many posts from specific segment of users (such as specific age grpup, specific nationalit, or just consumer vs. business)? It might be that we get ton of video posts but not picture posts. (good question. Mainly posts from consumers).
Me: That's a great piece of information that I would stash in my bank in addition to it's a worldwide drop. (ok, where do you want to go with diagnosis?)
Me: I want to evalute aspects I had mentioned before. Any code changes? (no). UI changes? (no).
Me: So it's worldwide event, not specific to devices but mainly specific to consumer posts but no patterns of demographics meaning it's impacting all consumers with some level of variances here and there but nothing too material. Ok, let's evaluate if there is news of some sort that;'s impacting us here. I honestly highly doubt here since this has been going on for about a year now you said and I would pressume if the news was year ago we would have taken care of this by now for sure. (correct).
Me: Ok, since you said year ago. Since when have we been noticiing change specifically? Do we have month, date, etc. that we know? (yes, it's been since April 2020).
Me: Great! That's critical data point. Considering it's worldwide event, no specific devices, consumer group but no specific demographics and it's happening since April 2020, it could be COVID related. Consumers mostly posts pictures of being outside (nice scenary, food pics, etc.). Well, the world pretty much shut down and people were cooked up inside home, didn't meet friends, etc. so there were no occasions to take pictures of other than with your kids (him or her doing funny things). I would expect such drop and maybe expect high usage of advertisement clicks. So, have we expereinced drops in CTR or CPI? (no those have gone up. In fact, shops feature has gone up too). There! I want to hang my hat on COVId being the issue. If that is the case, there is nothing for us to fix for now. I do expect we will start seeing some improvement as the economy opens up.
Interviewer: You are correct. Economy is opening up and I am seeing 2% increase now.
Me: Great! So if you are in agreement, and it seems like you are, there isn't much for us to solve here but we should keep an eye on these metrics jjust in case somethign else is at hand here.
Interviewer: Yes, I think you are right. Thank you for your time.
As I have a little idea of Instagram, I will use Facebook as an example here.
Redefining the problem
Why there is a 20% drop of daily active users of Facebook
Solutioning
Clarifying question - I would like to ask if this drop is abrupt i.e. happen over a period of few days or gradual i.e. over a period of months - Assuming it is an abrupt drop
The purpose of asking the nature of the drop significantly changes the direction of the investigation
1. Abrupt drop in the DAU
1.1 Segment by Demographics
1.2 Segment by lifestyle characterstics
1.3 Segment by user behaviour ( how they interact with product )
I will further divide each segment further
1.1. Segment by demographics
1.1.1 By Geographics
Hypothesis - Something has gone wrong like internet outage, government blocked the app) with particular region of the world
1.1.2 By users Age
Hypothesis - Govenment has blocked the access to the app for anyone below the certain age.
1.2 Segment by lifestyle
1.2.1 By users common interest
Hypothesis - Due to user clashes, users following a particular interest or topic stop coming to the platform
1.2.2 By following a known figure
Hypothesis - Due to some issue with the known figure ( he left the platform or appeal to its followers to leave the platform ) has triggered the chain of low DAU
1.3 Segment by behaviour
1.3.1 Segment by Device
Hypothesis - Due to recent changes in the mobile, users are struggling using the app and hence low DAU
1.3.2 Segment by usage
1.3.2.1 Segment by Power users
Hypothesis - Due to offline or online campaign - Power users are boycotting the platform
1.3.2.2 Segment by regular users
Hypothesis - Due to changes in platform policies and business model, regulars users are facing issues
Based on the model above, some questions to the interviewer. Basically these are data questions - If data was available I would do these analyses myself.
1. Do we have any recent changes in the policies which may impact the platform usage?
Interviewer - NO
Amit - then we can prune the 1.3.2 i.e. Segment by usage.
2. Do we have made any changes or roll out any changes on the mobile app or web?
Interviwer - NO
Amit - Then we can prune the 1.3.1 i.e. Segment by Device. As there are no changes recently in the app or web then it is highly unlikely that DAU has declined due to tech issues
3. Does data show, users who are not coming as much as they were coming before on the platform are of a particular age?
Interviwer - No, It is a mix age users who are not coming on the plaform as they were coming earlier
Amit - Ok then we can prune the 1.1.2 i.e. Segment by Age -
Now we are left with
1. Segment by Country
2. Segment by a common topic
3. Segment by following a known figure
Question ) Does data shows, if these users are from the particular country ?
Interviwer - Yes, they are mostly from India
Question ) Does data shows, if all these users following a common topic?
Interviewer - No , there are mix match of the topics they were following
Question ) Does data shows, if all these users following a common known figure?
Interviewer - Yes, they were following Actor John Miller
Conclusion - Due to the announcement from John Miller of taking off from Facebook, his followers are not coming on the platform as much as they were coming earlier. Hence affecting the DAU .
Top Meta (Facebook) interview questions
- What is your favorite product? Why?89 answers | 263k views
- How would you design a bicycle renting app for tourists?62 answers | 82.5k views
- Build a product to buy and sell antiques.54 answers | 66.8k views
- See Meta (Facebook) PM Interview Questions
Top Problem Solving interview questions
- A metric for a video streaming service dropped by 80%. What do you do?50 answers | 135k views
- You launched a new signup flow to encourage new users to add more profile information. A/B test results indicate that the % of people that added more information increased by 8%. However, 7 day retention decreased by 2%. What do you do?29 answers | 28.8k views
- Drivers are dropping out of a city on Lyft. How do you figure out what's going on?23 answers | 18.8k views
- See Problem Solving PM Interview Questions
Top Problem Solving interview questions
- Your new feature boosts Amazon Search by 10%, adds 2s to load time. What do you do?19 answers | 36k views
- There is a 15% drop in the open rate of Instagram App. You are the PM. Tell us what could have happened.11 answers | 10.1k views
- There is a data point that indicates that there are more Uber drop-offs at the airport than pick-ups from the airport. Why is this the case and what would you do within the product to change that?10 answers | 22k views
- See Problem Solving PM Interview Questions