How would you decide between showing more ads on the Facebook Newsfeed vs showing a "People you may know" recommendation widget?

I'd like to start by confirming my knowledge of the product, users, and goals

• FB newsfeed is essentially the "homepage" or the main screen of the FB app
• FB newsfeed is a scrollable timeline of content. The content is written posts, photos, and videos primarily.
• The conten is mostly from accounts that a user follows. It also can have promoted posts or ads that may be relevant to the user.
• Users of Newsfeed
  • Posters - individual accounts and buinsesses who create content and "post" to the newsfeed 
  • Viewers - those who look at the content in the newsfeed.
• While on the Newsfeed, viewers can take the following user actions:
  • likes, comments, shares, clicking (on an account or on a CTA)

• Viewers: find something on their newsfeed that is valuable (timely, important, entertaining and/or actionable content)
• Posters: others interact with their content and provide meaningful engagment with it 

• Amount and type of content on the newsfeed
• Segment engagements by type of content engaged with
• Time on the platform - this is mostly insightful if engagments drop, telling us if its a retention problem or a user behavior change
• Ads Metrics from Newsfeed - we assume # of engagements is correlated with ad clicks/conversions but want to monitor for anomalies or thresholds in that assumption.
  • CTR
  • Conversion rate

• Hypothesis 1:  As we increase RELEVANT ads (to a point), "mature" users will engage MORE, but "new" users who just joined will engage less
• Hypothesis 2:  If we add a "people you may know" widget in the newsfeed, particularly "new" users would click on and follow more accounts. For newer users, more followed accounts has a larger effect on time spent on platform and therefore # of engagements. So we'd see a gradual, but significant, increase in engagements over time. Whereas I don't think we'd see significantly more engagement from "mature" users when provided a "people you may know" widget.

• Test # 1
  • I'd pick a sample set of users and segment them into 3 test groups, new users, moderate users, highly active users
  • I'd have 5 levels of ad increases (e.g., x # of ads per y amount of content scrolled) 
  • I'd test all 5 levels randomly the same amount of times across several months and compare the # of engagements, time spent on platform, and ad metrics across the 5 levels.
• Test #2 
  • I'd pick a sample set of users and segment them into 3 test groups, new users, moderate users, highly active users
  • I'd pick a sample set of users who could be considered "newer", I'd work with data science to determine a good time period and # of followers that represents this group such as those who joined within the last 12 months and have less than 100 followed accounts.
  • I'd split the sample set in half--the experiemental and control group.
  • I'd then roll out the "people you may know" widget with the experiemental group for 3 months. This time period needs to be long enough to not just see first order effects, but also second order effects of following more people and how that increase in content on their newsfeed may impact their # of engagements

1.  
   1. Clarifications:
      1. Basically overhere we are trying to understand that shd the algo be modified to give high importance to ads or 'people you many know' widget in news feed.
      2. People you may now widget --> is friend suggestion widget right? where facebook shows recommendations on users whom the user can add as friends.
      3. Is it for some specific segment of customers or we are planning to make this change across facebook platform for all users?
   2. Any goal/objective which I should keep in mind due to which we are evaluating this choice? U can choose. ok so facebook's mission is 'Give people power to build communities and bring the world closer' and thus facebook needs to promote ways by which users can find people with similar interests thus building communities and should promote engagement within those communities thus helping worl to come close together.
   3. What is the satus co? as in currently what is shown to the users?
   4. So facebook is the social media platform........ News feed is the home page where users see updates/posts from friends, groups, pages they have liked, ads from brands which seem to match user interests and promoted widgets like 'People you may now', reel & short videos. Users scroll through their feed and engage with the posts by reacting, commenting, sharing and thus interact with their network (including friends, family & brands which they like) who have similar interests. 
   1. USer groups on facebook
      1. Content creators
         1. Casual users
         2. Brands/Influensers
      2. Content consumers
         1. New user 
         2. Casual users
         3. Brands/Influensers
   2. I can think of below attributes which we should consider for evaluating the options:-
      1. Engagement: Facebook's primary goal is to promote engagement with in the users via various features like posts with diff media types, reactions, comments, shares. But for engagement it is important that first users can find other users with similar interests which is where 'People you may know' feature helps. It recommends friends to users based on matching interests, mutual friends. Once people add friends then only they would engage with them. Prioritizing ads over friend widget is conflicting with the main goal of the platform.
      2. User experience for finding friends : if ads given priority over finding friends then the user exp for finding friends will face friction. Also usually this widget will be used by new users -> finding friends plays imp role in growing these users to medium freq users. If they cannot easily find their friends then they would quickly loose interest in the product itself as they won't receive posts from their friends thus won't be able to build their network with which they can engage.
      3. UX for consumers looking for engaging posts: From a user who aready has some 200+ friends on facebook, finding friends is not the first thing which he/she wants to see on their news feed. They have come on facebook to write posts, check reactions/comments on their posts, search for engaging content & ads from brands they love. If friends widget is put first then ads get pushed to 2nd fold of the page, considering 90% users are scrolling via mobile devices.
      4. Impressions for ads : If ads are given lower priority then ads would have lower impressions leading to lower CTR and thus the CAC of brands running ads would go down and thus spoil their experience. Ads are also imp as that is the main revenue driver for facebook. Considering many competitive platforms are coming up, spoiling brand experience might cost facebook with its revenue...
      5. Engg effort:- 
         1. Whether to show ads before friends widget or vice versa, seems to be a small tweak in the algo. As per my understanding facebook currently also has such tweaks for other widgets like reeels/videos, stories thus engg effort is small. 
         2. Also the capability to tweak algo based on user segment is also something which would exist thus engg effort erspective is low for this change.
      6. Migration: Suddenly customers will stop seeing friends widget.. many users won't even know that friends can be found via the menu --> friends --> click suggestions. So no. of connections made can be impacted. We would need to plan for a way to make them aware as to now how they can make friends.
   3. Based on these attributes, I would like to recommend to :-
      1. Show the 'People you may know' widget as it aligns with the company goal.
      2. Use impressions to switch between 'ads' and the widget. Where the impressions/week for widget will be based on user segment -
         1. If it is a new user (created account in last 30 days and has < 30 friends)
            1. Show Widget for 4 impressions/week and rest u can show ads.
         2. If it is an existing user
            1. Show widget for 1 impression/week and rest of the times u can show ads above the widget.
      3. Such an approach will ensure the company goal is met, revenue is not harshly impacted and experience of users is also not compromised.
   4. Based on the above recommendation would like to run A/B test to understand the impact of the hypothesis 
      1. Showing ads will impact no. of connections and thus engagement for new users
      2. Showing widget above ads would impact engagement of existing users

Goals

Facebook’s mission is to empower users to build community and a sense of belonging.

Newsfeed is the first product that most users see when they sign-in to Facebook. This is where they view and engage with stories from their friends / network / community as well as see paid stories from advertisers. 

Lets try and outline the goals for the NewsFeed product into business and user goals. 

User Goals

• View updates from network
• Share updates with network / build community. 
• Build network / Connect with others

Business Goals

• Show compelling and relevant content
• Grow time spent on Facebook
• Monetize users / Grow advertising business

To summarize user want to Connect with community and Facebook want to monetize users. 

To enable both a common goal would be increase engagement with the newsfeed. For users this comes in the form of compelling stories that they discover and people they can connect with. For Facebook this comes in form of the ability to make revenue through paid stories. 

User Actions

I would break down Facebook users into three broad buckets

• Active users —>  visit Facebook everyday and have 1000+ connections 
• Moderate users —> Casual users. May use the product 1x per week and have a moderate level of connections 
• Passive users —> visit perhaps 1x month. Have a handful of connections of close friends and family

Each of the above users had different patters of engagement with Facebook and the decisions on whether to show ads or PYMK should vary based on the user type. To inform our decision with data lets look at the key metrics that we want to track

Metrics

North Star 

My North start metric for this would be Engagement as it aligns closely with the goal.We want to uphold user engagement on the newsfeed because if there is a drop on that it reduces our ability to monetize the users.

To that end the key metric I would track is 

• Time spent in Newsfeed —> How do the changes we make impact the time spent on the newsfeed for existing users

Other Metrics

• Engagement with PYMK feature by user segment  —> Which user segment is engaging with this feature the most / adding connections? If one particular segment is engaging with the feature more then we should show it more to them rather than ads as growing a network increases the number of organic stories one would see which in turn reduces the ad load.

Counter Metric

We want to ensure that we dont turn people off by showing them too many ads or too many PYMK recommendations based on which segment they belong to. 

To track that I would measure retention

• DAU/MAU  —> Are we seeing a drop in the number of users logging into the app

Summary

To decide whether we want to show more ads or PYMK to users we need to segment our users into different groups based on the current engagement patters and their network. Thereafter based on the data around engagement and retention we can decide which route to take. This would be personalized offering and would vary from user to user as there is no one size fits all decision for this.

To start, I would first clarify this execution question. Is there any business goal we're trying to achieve? [Answer] no. Is there an option to show both features? [Answer] we're trying to decide which one to show at the top of the newsfeed. 

Okay, first I'd like to visit Facebook's goals. Facebook is a social networking platform that is aiming to connect people from around the world. Knowing this, I understand that Facebook relies heavily on how many customers sign up and use the platform. However, with this question I have to decide which one to show at the top. 

Let me think through the pros/cons of each feature 

Now, I'm going to segment user groups/types to understand who's using Facebook and why. The groups I will use are: high frequent users, casual users, and new users. 

High frequent users - I'm going to assume that this group of users already has a mature network established. Friend suggestions may/may not be relevant. They tend to use Facebook to find events, create events, catch up with old friends or browse content. They sign in multiple times a week. 

Casual users - this group of user is contining to build their network. Sometimes when they see friend suggestions, they view the persons profile and maybe make a connection. They use Facebook whenever they have to connect with someone or would like to browse content. They sign in multiple times a month. 

New users - this group is looking to established their network. They recently joined Facebook to see what it has to offer. 

If I think about a new users vs. a high frequent user - they have different needs from Facebook. For example, if we show a new user ads instead of friend suggestions - that will defeat the purpose of Facebook (which is to connect people from around the world). However, for a frequent user they have their network estabilished so showing relevent ads will help drive engagement. 

Since, engagement and retention is important for Facebook - I would run an A/B test to see which feature impacts engagement and retention. Some metrics I would track would be: click through rate (CTR) for ads and number of connections made through PYMK feature. 

Ultimately, if no one is using Facebook (and the retention is low) Facebook will not have oppertunities to show ads. Therefore, users building their social network is important and having retained users so that allows more oppertunities to show ads.