Graph Search was a semantic search engine launched by Facebook in 2013 to allow users to perform more natural, conversational queries using real-world connections. Unlike traditional keyword searches, Graph Search leveraged Facebook’s social graph—users, likes, check-ins, and interests—to return contextually rich results.
What Graph Search Could Do
Example queries included:
- “Friends who live in New York and like sushi.”
- “Photos I liked in 2014.”
- “Friends of friends who work at Google.”
It allowed highly specific search filters based on relationships and behaviors—data that only Facebook had access to.
Applications & Limitations
For businesses, Graph Search initially seemed promising for:
- Recruitment (find people with certain job titles in regions)
- Audience research
- Content discovery
However, privacy concerns and limited adoption led Facebook to scale it back. As of now, Graph Search has been deprecated and replaced with more keyword-based search.
Legacy and Lessons
Graph Search was an early experiment in semantic search—an effort to understand user intent, not just keywords. Today, similar concepts are applied in Google’s BERT and AI-based algorithms, and Facebook continues to use structured data for internal ranking and discovery, even if the feature isn’t publicly branded.
