Facebook’s “People You May Know” feature suggests new friends based on several factors. The algorithm looks at mutual friends, networks, workplaces, schools, and other connections to recommend potential friends. Some common reasons a person may show up as a suggestion include:
You Have Mutual Friends
If you and another person have several friends in common, Facebook will likely suggest them. Facebook’s algorithm computes how many friends you share and the strength of those connections to determine relevance. More mutual close friends make it more likely someone will be recommended.
You’re in the Same Network
Facebook allows users to list networks they are part of, like their school or workplace. If you and another person are in the same networks, the algorithm sees this as a strong connection and will likely recommend them. Fellow alumni, coworkers, classmates, and other network connections often appear as suggestions.
You Have Interacted Before
If you and another person are not directly connected but have interacted before, such as commenting on the same post, liking each other’s content, or attending the same events, Facebook’s algorithm picks up on these signals. Previous interactions suggest you may know each other and want to connect.
You Have Similar Interests
Facebook analyzes information you provide in your profile, pages you’ve liked, posts you’ve engaged with, and other activities to determine your interests. If you and another person have similar interests, you are more likely to appear as mutual friend suggestions. Common interests suggest you may get along well.
You Are in Similar Demographics
Location, age, and gender can play a role in People You May Know suggestions. People who are in your age group or live nearby are more likely to run in similar social circles. Facebook may analyze demographics like these as part of generating recommendations.
How Facebook’s Algorithm Works
Facebook’s friend recommendation algorithm considers hundreds of signals to suggest new connections. Some key factors the algorithm analyzes include:
Mutual Friends
As mentioned, the number of friends you share with another person is a strong indicator of relevance. More mutual friends means a higher likelihood of being suggested.
Strength of Connections
The algorithm not only looks at how many mutual friends you have, but also the strength of those connections based on interactions. If mutual friends frequently interact with both you and the suggested person, that’s a stronger signal.
Network Overlap
Shared networks, especially smaller networks representing schools, companies, or other defined groups, significantly boost the likelihood of recommendation.
Interactions
Actions like liking, commenting, tagging, and messaging back and forth increase the probability someone will get recommended, even without shared connections.
Interests and Traits
Common interests, habits, pages liked, groups joined, and other traits determined by Facebook’s algorithms can factor into People You May Know.
Location
People who live in your area or visit frequently are more likely to be suggested than those far away, since proximity increases the chance of real-life connections.
Demographics
Age, gender, workplace, schooling, and other demographic traits might play a role in suggestions.
Partner Targeting
Advertisers and apps on Facebook can opt to have People You May Know recommendations dynamically generated based on targeting criteria.
Limits of Facebook’s Algorithm
While Facebook’s algorithms are sophisticated, they have flaws and limitations, including:
- Bias – Formulas can recreate biases if the training data contains imbalanced demographics or stereotypes.
- Limited signals – Not all relevant variables can be analyzed, like real-world interactions outside of Facebook.
- Gaming the system – People can intentionally manipulate signals like likes and comments to get suggested to others.
- Imperfect connections – Just because two people have traits in common doesn’t mean they will get along or want to connect.
- Creepy effect – Suggestions based on limited interactions can feel overly invasive and lead to rejection.
The technology has improved over time but still lacks human insight and intuition.
Reasons a Suggestion May Feel Unexpected
In some cases, a People You May Know recommendation can seem odd or random. Here are some possible reasons you may find yourself scratching your head:
You Have a Mutual Connection You Forgot About
You likely have forgotten connections from your past that link you to others. Even one mutual Facebook friend from years ago can cause someone to get suggested.
One Side Interacted More
If one person liked, commented on, or browsed the other’s profile more frequently, it can create a one-sided recommendation. The levels of interest and interaction may be different.
You Have Hidden Traits in Common
Facebook’s algorithms dive deep, analyzing many less visible signals beyond your profile. Shared hidden traits like browsing patterns, locations visited, and interests can cause surprising suggestions.
The Algorithm Made a Mistake
No algorithm is perfect. Occasionally the formula misfires and suggests someone with few real connections. Garbage in, garbage out.
More Distance Than Expected
People You May Know can surface connections beyond your extended social circle, which can feel jarring. Suggestions aren’t limited solely to close friends of friends.
You Have Hidden Connections
In some cases, you and the suggested person may share real-world connections you are unaware of, like extended family, neighbors, old classmates, former coworkers, or other links.
How to Manage Suggestions
If unwanted people appear in People You May Know, you have options:
Remove a Suggestion
Click the “X” on a suggestion to remove them. This provides feedback to Facebook’s algorithms. However, the person may reappear later if signals indicate a strong connection.
Restrict Your Profile
Limiting who can see your friends list and posts prevents irrelevant people from accessing data that may get you suggested. Tighten privacy settings if unwanted recommendations persist.
Disconnect From Mutual Friends
Unfriend or restrict mutual friends who may be linking you to the unwanted person. Breaking shared connections helps avoid recommendations.
Contact the Person
Politely ask the recommended person to remove you as a follower or friend if they have connected. This takes you off their list of connections.
Report Suspicious Behavior
If you believe someone is inappropriately manipulating recommendations, report them to Facebook so corrective action can be taken.
Turn Off Recommendations
As a last resort, you can disable People You May Know suggestions entirely in the app’s settings. However, this prevents potentially relevant recommendations too.
Conclusion
Facebook’s friend recommendation algorithms utilize complex machine learning techniques. While not perfect, the formulas analyze numerous signals to surface potential connections. Some suggestions may seem random but often have logical explanations. Carefully managing privacy settings, connections, and feedback to Facebook can help refine People You May Know and keep recommendations relevant. With billions of diverse users, crafting the ideal formula remains an ongoing challenge.