There has been a 15% drop in usage of Facebook groups. How would you fix it?

Share Knowledge
- Feel Intelligent
- Feel Satisfied
Share Viewpoint
- Feel connected with peers
- Feel connected to cause
- Feel part of group
Ask Help
- Looking for information
Market Product
- Sell Something
- Market their product or services

For this question let's concentrate on "Sharing Viewpoint" as most common motivation behind users posting content on Facebook Groups.

User Journey:

User likes a topic.
User searches for a group.
User joins the group.
User reads/like/comments on posts.
User makes his first post.
Other users read/like/comment on his post.
User engages with other like minded users.

User Pain Points(Probable):

Users don't feel a need to post anything.
- Concept of community doesn't exist anymore
Users don't view groups as viable options for posting anything.
- They find other channels more viable
- Advertisers crowd the group
User find difficult to post something in groups.
- Stricter policy
- UI not supportive
- Privacy issues
Users don't get what they seek by posting in groups.
- Other users don't engage (Most Likely)
- Community is getting weaker

Considering personal experience I would select last pain point as a reason for reduction in posts.

Problem Description: Users are not sharing their view points in Facebook groups because other users don't engage with them anymore.

Redefined Problem: Improve user engagement in Facebook groups so that users are motivated to post their view points in groups.

Solutions:

Tweak Feed algorithm to prioritize group posts
Autopost most engaging posts across similar groups
Personalised recommendation of groups to users
Group Chat feed

	Tweak Feed algorithm	Autopost across Groups	Personalised Recommendation	Group Chat Feed
Value to User	M	L	M	H
Value to Platform	M	L	L	H
Resource Investment	L	L	L	M
Time Investment	L	L	L	M

Considering drop of 15% in usage metric, I would recomment implementing Group Chat functionality for groups. User can participate in live chat in groups. This will lead to high engagement and improved motivation for users to post content leading to improved usage metrics.

Metrics:

Business
- Post/Group/User
Product
- Chat Sessions
- User/chat sessions
- NPS
Anciliary
- Platform Metrics

srocks · Answer 2 · 2019-04-27T20:29:04+0000

15% drop in FB groups. FB Groups is a feature where people discuss, share or connect over different topics. Does this 15% drop in engagement? or a 15% drop in new users or a 15% drop in groups created.

Assume its engagement drop here i.e. sessions per day have dropped in FB groups - which means people engage lesser with the groups day to day than before.

I have a couple of clarification questions here:

- is the drop recent or has happened over - lets say it has happened over 1 month now

- did the drop happen on android, ios, web? if web then a particular browser?

- any particular geography

- changes in external ecosystem- competitors, data plans from n/w providers, etc.

- did we release a new feature that led to the drawing of users from FB groups

- was there a technical glitch

- seen a drop in related feature such as pages, feed, etc.

If the answer to all above is no then the drop is mostly due to behavioural than technical reasons.

- Metrics- session per day, actions per session, % engaged users, median new posts per day in a group, median responses per post

Lets say median new posts per day in a group have dropped that have led reduced engagement. I would then go about understanding the drop in this particular metric and see what are the possible reasons:

- seasonality

- drop in particular kind of groups e.g. schools related or education related communities because exams are there

- Notification bug - other people are notified when a post is created and thus response rate is low

- check churn rate and if that has increased talk to churned users to see reasons of churn

PM Team · Answer 3 · 2018-08-10T05:48:27+0000

First, you would clarify if this question is primarily about root-cause analysis, i.e. identifying the reason for the 15% drop.

Assumption: Facebook Groups are defined as a group of Facebook members that share a common interest, organization, family, etc.

Questions to ask to root-cause the problem:

Basic questions
a. Is the 15% drop YoY or over the projected growth?
b. Is the 15% drop for a specific geographic region? Or across the world?
c. Is the 15% drop for a specific demographic group?
d. Did we see a drop in DAU, MAU in the same period?

Did the 4Ps change?
P1 - Did we change the product is anyway that would cause the drop? Is the product still relevant? Did we change any adv. algorithm?
P2 - Did we change any advertising / marketing campaigns that would cause the drop? (Negative emotions)
P3 - Did we change the placements of ads?
P4 - Did internet providers or mobile providers change pricing? Or costs increase for users?

Others potential issues -
Were there any outages with data centers? Or servers?
Did the users start using a competitors / substitute product ?

Other Parameters:
If the drop is regional
- were there any weather events or acts-of-god during the drop period?
- were there any political events?
- were there religious / cultural / seasonal events?

Once you have identified the root cause you could take specific actions to fix it -
For example, you might realize that the advertising algorithm changed and displaying too many ads within groups. We could fix that.
For example, you might realize that heavy usage of videos/ads could have caused overage charges thereby increasing monthly cost. We could introduce a "low bandwidth mode" for mobile users.

tisvesalem · Answer 4 · 2019-09-30T16:56:47+0000

First I would do a root cause analysis (Internal vs External)

Internal

External

Is this an L1, L2, or L3 metric?

Did a total # (denominator) of the metric drop or did a numerator metric drop?

Did critical sources go down? Server? Deployment

What is the timeline for this drop? Did it happen in a day? Did it happen in a month?

What changes occurred? Did a new feature come out? Did a new product come out? Has there been a code change? Did the model algorithm change?

Are there any telemetry clues? Any errors or checkpoints?

Is the independent metric affected, or all the related metrics as well?

Is it a seasonal issue?

Particular geography?

Has there been bad PR?

Has there been turbulence in the political sentiment?

Has there been a self-product launch for cannibalization?

Has there been a competition launch?

Has our stock plummeted?

If it is an internal bug, after root causing the metric, a bug would be filed with a severity level to

1. Data team

2. Feature dev team

3. Model Research Team

4. Critical resources team

15% is a pretty heavy drop in an L1 metric, so, depending on the significance of the metric to the overall objective and key results of the org, the sev on it would change, and the timeline for the fix.

For an external issue, it would be better to identify mitigation work-around strategy, set a timeline for when you would think things would bounce back, and work in parallel.