You're scrolling through Facebook at 2:00 AM. Suddenly, a face pops up in that middle-of-the-feed carousel. You don't know them. You check the profile—nothing. There are zero shared connections. It’s a total stranger from three towns over or maybe someone you haven't thought about since third grade. Seeing Facebook people you may know no mutual friends is honestly one of the creepier parts of the modern internet experience. It feels like the algorithm is watching you through your window.
But it isn't magic. It's math.
Facebook’s "People You May Know" (PYMK) feature is essentially a massive prediction engine. While mutual friends are the easiest way for the system to link people, they aren't the only data points available. The company has spent nearly two decades refining how it maps the "Social Graph." This graph isn't just about who you've added; it’s about where you go, who you might have emailed, and whose profile you accidentally clicked on during a late-night rabbit hole.
The Data Behind the "No Mutual Friends" Mystery
Most people assume Facebook only looks at your current friend list. Wrong. The algorithm is way hungrier than that. When you see Facebook people you may know no mutual friends, the system is likely pulling from "shadow" data or secondary signals.
Think about your phone contacts. If you’ve ever given the Facebook or Messenger app permission to access your contact list, you’ve fed the beast. Even if you don't have mutual friends with "Sarah from Accounting," if her phone number is in your contacts and your number is in hers, the algorithm connects the dots. It doesn't care if you haven't spoken in five years. The link exists in the metadata.
Then there’s the "Network Proximity" factor. This is basically the digital version of six degrees of separation. Maybe you don't have mutual friends with someone, but you share a specific, niche interest. You’re both in the same hyper-local "Community Gardening" group. Or perhaps you both attended the same university during the same years. Facebook’s algorithm, which uses a combination of machine learning models and large-scale graph processing, identifies these clusters. It figures that if you inhabit the same digital spaces, you’re likely to know each other in the physical world.
Why Your Location Might Be Snitching on You
Location services are a huge part of this. While Facebook has occasionally been vague about exactly how much "real-time" GPS data influences PYMK, technical patents and user patterns suggest it plays a role. If two people spend a significant amount of time in the same precise GPS coordinates—like a small office building or a specific gym—the algorithm notices.
📖 Related: Why the CH 46E Sea Knight Helicopter Refused to Quit
It’s about "Co-location."
If you're wondering about Facebook people you may know no mutual friends, ask yourself if you’ve been anywhere specific lately. Did you go to a wedding? A conference? A dive bar? If you and another person were both there, and you both have location services toggled on, the algorithm might decide you’re worth a suggestion. It’s trying to facilitate that "Oh hey, I saw you there!" moment, even if it feels a bit like stalking.
The Creepy Reality of Profile Viewing
We’ve all heard the rumor: "If they show up in People You May Know, it means they’ve been looking at your profile."
Facebook officially denies this. They’ve stated multiple times that profile views do not influence these suggestions. But the internet is skeptical. Why? Because the "black box" of the algorithm is notoriously opaque. While "looking at a profile" might not be a direct "if-then" trigger, the interaction data is definitely tracked. If you search for someone’s name—even if you don't click "Add Friend"—you’ve signaled interest.
The algorithm prioritizes engagement. It wants you to stay on the platform. If showing you a "stranger" (who actually happens to be your ex’s new partner) keeps you clicking and scrolling, the algorithm has technically succeeded, even if it made you feel a bit sick.
Digital Footprints and Third-Party Data
Another layer involves the information Facebook buys or tracks outside of its own app. Using the Meta Pixel—that tiny bit of code on millions of websites—Facebook knows where you shop and what you read.
👉 See also: What Does Geodesic Mean? The Math Behind Straight Lines on a Curvy Planet
Let's say you and a stranger both frequently visit a niche hobbyist website for vintage Porsche parts. You don't know each other. You have no mutual friends. But the Meta Pixel has tagged both of you as "Porsche Enthusiasts." When the PYMK algorithm looks for "highly relevant" people to show you, it might bridge that gap. It’s not just about who you know; it’s about who you are according to your data.
How to Take Back Your Privacy
If seeing Facebook people you may know no mutual friends makes you uncomfortable, you aren't stuck with it. You can't turn the feature off entirely—Facebook really wants those connections to happen—but you can muzzle it.
First, go into your "Apps and Websites" settings. Look at what you've shared. More importantly, go to your "Permission" settings and find "Upload Contacts." If this is on, Facebook is constantly scanning your phone book. Turn it off. Delete the contacts you’ve already uploaded. This usually results in a drastic change in who the algorithm suggests.
Next, manage your location settings. On your phone (iOS or Android), you can set Facebook’s location access to "Never" or "Only While Using the App." This prevents the background tracking that leads to those "We were at the same coffee shop" suggestions.
Practical Steps to Clean Up Your Suggestions
- Purge Uploaded Contacts: Head to the "Manage Contacts" page in your Facebook settings. Hit "Delete All." It won't happen instantly, but the system will eventually stop using that data.
- Adjust Privacy Settings: Change who can send you friend requests. If you set this to "Friends of Friends," it limits the pool of people the algorithm can suggest to you.
- Stop the Search: If you’re "creeper searching" people, stop. The algorithm uses your search history to refine its suggestions.
- Click the 'X': When a stranger pops up in the PYMK list, click the small 'X' in the corner. This provides "negative feedback" to the machine learning model. It tells the system, "I don't know this person and I don't want to." Over time, it gets the hint.
What Most People Get Wrong About the Algorithm
The biggest misconception is that the algorithm is "broken" when it shows someone with no mutual friends. Actually, it's working exactly as intended. Facebook’s goal is "network expansion." If it only showed you people you already have 20 mutual friends with, the network would stay stagnant. By introducing "weak ties"—those people with no mutual friends but shared backgrounds or locations—Facebook expands your social web.
Sociologists call these "weak ties" incredibly valuable for job hunting and information spreading. Facebook just calls them good for business.
✨ Don't miss: Starliner and Beyond: What Really Happens When Astronauts Get Trapped in Space
It’s also worth noting that "No Mutual Friends" might just be a privacy setting. The person being suggested might actually have 50 friends in common with you, but they’ve set their friend list to "Private." You can't see the connection, but the algorithm can. It knows you're in the same circle; it's just not allowed to show you the proof.
Looking Forward: The Future of Suggestions
As AI and machine learning become more sophisticated, these suggestions will likely get even more specific. We’re moving past simple location and contact matching. We’re entering an era where behavioral patterns—how long you hover over a photo, the sentiment of your comments, the types of videos you watch—will dictate who ends up in your "People You May Know" list.
The feeling of being "watched" isn't going away. However, understanding that these suggestions are based on cold data points rather than a literal spy in your room makes it a bit easier to manage. It's just a computer trying to find patterns in the noise.
To stop the cycle, you have to starve the algorithm of data. Minimize the permissions you give the app. Be intentional about your privacy settings. And remember, just because someone shows up in your feed doesn't mean you have to acknowledge them. Sometimes, the best way to handle a weird suggestion is to just keep scrolling.
Immediate Action Items
- Check your Facebook Audience and Visibility settings right now.
- Check "How People Find and Contact You" and restrict friend requests.
- Review your Off-Facebook Activity to see which apps are sending your data back to Meta.
- Disconnect any third-party apps that don't need access to your profile.
By tightening these digital bolts, you can turn that "creepy" suggestion list back into something that actually resembles people you might, you know, actually know.