Facebook Friend Suggestions: Finding, Ranking and Presenting People a User May Know
Company: Meta
Role: Product Manager
Category: Product Design & Strategy
Difficulty: medium
Interview Round: Other
Design the Facebook feature that suggests people a user may know and might want to add as a friend. You may approach it technically, for example by laying out decision-tree logic for which candidates qualify and in what order, or with a product design framework that works from users and their needs to features. Either way, explain whom the feature serves, where candidate suggestions come from, how they are ranked and shown, and how you would measure success.
```hint Two people per suggestion
Every suggestion involves the person who sees it and the person being suggested. Ask which offline relationships leave traces on Facebook, and which suggestions could feel intrusive or unsafe to either of them.
```
### Clarifying Questions
- Are we designing friend suggestions from scratch, or improving the version Facebook already offers, and if so, what problem prompted the work?
- Which users matter most: newcomers with almost no friends who need a network quickly, or established members whose friend list is mostly complete?
- What should success mean: more accepted friend requests, connections that lead to ongoing interaction, or overall network growth?
- Which data may matching use: phone contacts the user chooses to upload, school and work history on profiles, group membership, profile visits?
### What a Strong Answer Covers
- A goal tied to genuine connections rather than raw request volume, and a reasoned choice of user segment.
- Candidate sources based on evidence of real-world acquaintance, such as mutual friends, overlapping school or work history, shared groups, uploaded contacts and being tagged together, with logic or a model that orders them.
- The product experience: where suggestions appear, how each one is explained, and how users dismiss them or limit who can find them.
- Privacy and safety for both people, including sensitive inferences, minors and unwanted contact.
- Success metrics with guardrails, such as acceptance, interaction after connecting, dismissals, ignored requests, reports and blocks.
### Follow-up Questions
- Suggestions based on profile visits or uploaded contacts can reveal something a user wanted to keep private. Where would you draw the line?
- How would suggestions differ for someone who joined yesterday and someone with several hundred friends?
- How would you stop scammers or harassers from using the feature to find targets?
Overview: A product design question asking how to build the Facebook feature that suggests people a user may know, using either decision logic or a product framework. It tests choosing goals and users, finding and ranking candidates from real-world connection signals, protecting privacy and safety, and defining success metrics.