Facebook’s friend suggestion feature has been around since the beginning of the social network. The algorithms powering these recommendations have evolved over time, but friend suggestions remain a core part of the Facebook experience in 2023.
How Facebook friend suggestions work
Facebook uses a variety of factors to determine which friend suggestions to show you. Some of the key signals Facebook considers include:
- Friends of friends
- People who have lots of mutual friends with you
- People you may have met offline
- Contacts imported from your address book
- People who go to the same school or work at the same company
- People who have interacted with your posts or profile
- Location data
- Groups and events you have in common
- Pages you both follow or like
The friend suggestion algorithm uses machine learning models trained on the above signals and patterns from billions of connections on Facebook. The models try to predict which people you are likely to know or want to connect with based on the available data.
How the friend suggestion algorithm has evolved
Facebook’s algorithms and the data used for friend suggestions have changed significantly over the years as the company has adjusted to evolving user behavior, expanded its data collection, and improved its machine learning capabilities.
In the early days of Facebook, friend suggestions were primarily based on your school or workplace networks, mutual friends, and imported contacts. As the network grew and diversified, Facebook added more signals to the mix.
Some key evolutions include:
- 2012: Location data from mobile devices added to suggestions
- 2013: Open Graph actions (likes, shares, etc.) incorporated
- 2014: Additional machine learning models trained on network patterns
- 2016: Facial recognition from photos suggested friends who looked similar
- 2018: User interactions like comments and reactions added as signals
- 2020:Removing facial recognition and emphasizing common interests/connections
In addition to using more data signals, Facebook has focused on making the friend recommendation algorithms more efficient, personalized, and relevant over time through advances in machine learning and deep learning.
Controversies around friend suggestions
Facebook’s use of data to fuel friend suggestions has also sparked criticism and controversies at times. Some examples include:
- Suggesting friends of ex-partners after breakups
- Outing LGBTQ+ users by suggesting their contacts
- Privacy concerns around using shadow contact information without consent
- Suggesting harmful or dangerous contacts like bullies or abusers
- Bias issues from making assumptions based on limited data signals
In response to some of these issues, Facebook has added tools to block specific people from being suggested as friends and has removed controversial data signals like facial recognition.
Current state of Facebook friend suggestions
Even with the evolution and controversies, friend suggestions remain a major feature on Facebook today. Some key facts about their current status:
- More than 65% of Facebook friends are discovered through suggestions
- Over 1 billion friend connections are made per day on Facebook
- The average user has over 150 friend suggestions waiting at any time
- Friend suggestions are personalized for each user based on thousands of signals
- Machine learning models continually train on new data to improve relevance
So while the algorithms and data powering them have changed, friend suggestions are still very much front and center for Facebook users in 2023.
Where to find friend suggestions
There are a few main places friend suggestions surface on Facebook currently:
- The Friends tab
- “People You May Know” box on the homepage sidebar
- Notifications when you have new suggestions
- When viewing friends of friends profiles
- After importing contacts or joining groups/events
Facebook tries to proactively push the most relevant suggestions through notifications and homepage placement based on its algorithmic assessments.
Types of friend suggestions
While the specific suggestions shown are personalized, most suggestions tend to fall into a few general categories:
- Friends of friends: The classic suggestion type, trying to help you connect with second-degree connections.
- Old friends/acquaintances: People you may have lost touch with from your school, hometown, past jobs, etc.
- New contacts: People from new groups/events or contact imports who Facebook thinks you might know.
- Network expansion: Suggestions to help you expand your network by meeting new people.
Facebook tries to strike a balance between reinforcing existing bonds vs. helping users make new connections when ranking friend suggestions.
Why Facebook shows you certain friend suggestions
It can sometimes be puzzling why Facebook suggests certain people. Here are some common reasons you may see specific friend recommendations:
- You recently joined a new group or event together
- The person liked, commented on, or shared one of your posts
- They went to your school or worked at your company
- You have over 30 mutual friends in common
- Their profile indicates similar interests or geography
- They recently searched for you or viewed your profile
- One of your close friends is also their close friend
Essentially, Facebook is looking for strong points of connection or commonality that indicate you may actually know each other or want to connect.
How to manage friend suggestions
If you want to curate who gets suggested, there are settings to control friend recommendations:
- Go to Settings & Privacy > Settings > Privacy Settings
- Under “How people can find and contact you”, click Edit for “Who can suggest you as a friend”
- Choose from “Everyone”, “Friends of Friends”, or “No One”
You can also block specific people from being suggested again by clicking the “X” on their suggestion or by blocking them entirely.
Turning off friend suggestions
If you don’t want any friend suggestions at all, you can turn them off:
- Go to Settings & Privacy > Settings > Privacy Settings
- Under “How people can find and contact you”, toggle “Do you want us to suggest friends for you?” to Off
This will disable friend suggestions system-wide. You can always turn them back on later if you change your mind.
Other ways to find friends
If you don’t want to rely on Facebook’s algorithms, here are some other ways to find and connect with people:
- Search for names of specific people
- Use the Friends tab to browse friends lists of people you know
- Join interest-based groups and connect with members
- Interact with profiles and content from people you want to know
- Import contacts from your email, phone, or other networks
- Connect your other social media accounts like Instagram or LinkedIn
The future of friend suggestions
Facebook will likely continue refining its friend suggestion algorithms as long as the feature provides value to users.
Some potential changes on the horizon include:
- More personalized and contextual suggestions based on expanded user data
- Suggestions optimized not just for connecting but for meaningful interactions
- Explanations for why certain friends are suggested to increase transparency
- User controls over which suggestion signals are used
- After the discovery stage, also suggesting ways to deepen friendships
Facebook is investing heavily in AI research and applying advanced techniques like deep learning and reinforcement learning across its products. The sophistication of friend suggestions will likely keep pace with these evolving technologies.
However, privacy issues and harmful use cases like excluding marginalized groups will require ongoing vigilance. The algorithms will need to balance relevance and user control with ethics and social impact.
Conclusion
Despite some controversies, Facebook friend suggestions remain a core way that people discover new connections and strengthen existing bonds on the platform. The algorithms powering suggestions have grown vastly more advanced over the years even as the core function remains the same.
Facebook seems committed to continuing innovating and evolving friend recommendations going forward. But they will also need to empower users with more transparency, control, and oversight to avoid criticism of overreach or harm. If done responsibly, friend suggestions can likely provide social value for many years to come by harnessing the power of technology to bring people together.