It’s been less than a month since Facebook expanded Graph Search to include status updates and posts, and the feature still isn’t available to all users.
Nonetheless, Facebook search engineer Ashoat Tevosyan shared an “under the hood” look at how it works — and why it took so long to become reality.
There’s nothing incredibly newsy in the post, but it’s educational from a search perspective. Here’s a bullet list of facts and figures from the post:
- post search has been two years in the making, and began as a hackathon project when Tevosyan was still an intern
- there are a billion posts every day on Facebook, and the posts index currently has more than a trillion posts
- post data such as time, location and tags are stored differently in posts, check-ins and photos — making the search project more challenging
- at 700 terabytes, the posts index is the biggest of all Facebook search indexes
- Graph Search currently uses more than 100 ranking features to figure out the best match when doing post searches
There’s more in the full article that might interest you, and it’s not incredibly technical (which is good for readers like me!). Unfortunately, there’s no word on how the rollout is progressing.