Facebook’s Real Time Social Network Search Patent

Will Facebook someday launch their own search engine that enables you to search the Web? That question surfaces every so often, without any real definitive answers. It’s possible that they might someday, especially since they’ve been hiring a number of people with job experience from some of the major search engines.

The last few times I recall seeing the possibility of a Facebook search engine raised was when a couple of different Facebook patents originally acquired from Friendster were granted, and each described an aspect of how a search engine might use connections on a social network to influence results seen on a search of the Web. (For example, see this post: Facebook Patents “Curated Search” To Attack Google.) A different question, and one that is just as interesting is how the search on Facebook itself works. A patent application published at the USTPO today gives us some ideas of how it may work.

A screenshot of Facebook search results on the term [cincinnati reds] on a search of posts by everyone.

I have to confess that I’ve rarely searched at Facebook, and I don’t think that I’ve ever searched using anything other than the search box at the tops of pages. There is a Facebook search interface, as seen in the image above, that allows you to choose between the following to search for:

  • All Results
  • People
  • Pages
  • Groups
  • Apps
  • Events
  • Web Results
  • Posts by Friends
  • Posts by Everyone
  • Posts in Groups

The Web results are powered by Bing, and likely follow Bing’s ranking algorithms in determining what to show first. The Facebook patent application provides details on Facebook’s realtime search of Facebook posts, and gives us an idea of how they might be ranking those. It includes a number of ranking considerations that it describes in possibly deciding what order to show you results when you do a search . The patent application describes what some of those may be when you do a search for “Posts by Everyone”.

Level of connection – Results might be shown from people you are directly connected to first, then from indirect connections, then random users of the network, possibly followed up by third party content published outside of the social networking system.

Reputation of Users – Posts from users who have low reputations, possibly spammers or malicious users, may appear lower in results than posts from users with higher reputations.

Popularity – A popularity score for individual posters, based upon having more interactions than other users, may cause posts returned in a search to rank more highly than posts from less popular users.

Similarity – A similarity score between the post’s author and the searching user could be based upon demographic information such as age, gender, location, interests, or other similarity measure, and that similarity could also boosts results from similar posters.

Affinity with other users – if you interact with certain people more frequently than others, their relevant posts would rank more higly in Facebook’s search results.

Affinity with shared applications – for example, if you frequently participate in social online puzzle games, the posts from other people who participate in the same kind of puzzle games (on the social network itself, or possibly on the Web) may be boosted in search results. The patent filing considers this kind of shared affinity to be a “personalization” approach because it takes advantage of interests that are uncovered based upon your use of the social network or even on the Web, to personalize your results by ranking higher posts from other people who may share interests with you.

The patent filing provides much more in the way of details on how Facebook’s search index is set up, and is definitely worth a look.

Real Time Content Searching in Social Network
Invented by Akhil Wable, Hong Yan, Spencer Ahrens, Yofay Kari Lee, and Guizhen Yang
Assigned to Facebook
US Patent Application 20110196855
Published August 11, 2011
Filed: February 11, 2010


Indexing and retrieving real time content in a social networking system is disclosed. A user-term index includes user-term partitions, each user-term partition comprising temporal databases. As a post is received from a user, a user identifier, a post identifier, and a post is extracted. An object store communicatively coupled to a temporal database for recently received content is queried to determine whether terms in the post has already been stored. A term identifier is stored in the user-term index with the user and post identifiers. A forward index stores the post by post identifier.

Responsive to a search query, the user-term index is searched by the user’s connections and the terms. A real time search engine compiles the results of the user-term index query and retrieves the stored posts from the forward index. The search results may then be ranked and cached before presentation to the searching user.


When I start typing into the search box for Facebook, it shows a dropdown that lists pages and then people. I’ve assumed that’s all that it shows, and didn’t realize that it would show other types of search results in that dropdown as well, such as “posts by friends” and “posts by everyone.”

I find myself wondering how many people actually perform searches for “Posts by Friends” and “Posts by Everyone” or the Bing Web searches, and whether the site would be more interesting and useful if those types of results were more prominent. Should there be a “search” link in the main menu for Facebook pages, instead of or in addition to the search box?

Or, are the Facebook search features purposefully de-emphasized to keep people focusing upon the other features that Facebook offers?

19 thoughts on “Facebook’s Real Time Social Network Search Patent”

  1. interesting concept Bill,

    The ranking considerations highlighted by the model are similar to the approach we’ve taken in the past when considering influence amongst users of social networks.

    For social media monitoring where someone would be knowledgeable about a subject (i.e. they have previous posts including keywords associated with the topic), and would have a large number of friends or followers (also incorporating a follower to following rank), we would prioritise their considerations, requests or concerns above those of someone else.

    I can see a social media search function combining this and the other considerations providing some pretty valuable responses. I would suppose that applying a value against each of the different elements to give a truly individual response to a search query would be a pretty powerful tool too.

    I was thinking back to the example of the Levi’s Vintage Jeans example from a previous article and it would be great to get recommendations from people within your circle of connections, rankings based on geography and region from out of your circle with rankings based on the like button. For a consumer focused search it could be a pretty neat solution also benefitting from the instant interaction offered by facebook (someone could post a picture of those stone was levi’s!)

    I’m not sure on the reason why it’s tucked away at the moment – the strategist in me would suggest it was probably Facebook avoiding a ground war, but since google launched the Google Plus product – it’s probably because their waiting for the most opportune time to launch it!


  2. To Bill and the other readers,

    am I the only one that thinks Facebook’s search results are the poorest user experience? 50% of the time I look for something I can’t find what I was looking for.

    Sometimes I can’t even find 1million + business pages. Come on Facebook! Come on bing!

  3. Interesting stuff. Although you can patent search now? That’s a pretty worrying concept. Google+ is so similar to Facebook it shows how easy it is to get around these patents so all this really does it give away a load of trade secrets and plans for Google to pour over.

  4. If Facebook is ever going to compete in the search arena, they will not have the luxury of doing it in stealth fashion like this.

    They will have to produce a robust algorithm and attack Google “head on”, not because Google has an algorithm that can’t be reverse engineered, but because they would have to attack the mentality of humanity so as to be viewed as a search company. As it stands, FB is just social media and nothing more.

    As you stated, Bill…you never use Facebook for search…it goes without saying because, well, no one does…


  5. I really enjoy these kinds of “social search” posts Bill. As much as we all love Google, Facebook’s edgerank algo probably plays as much into content discovery as anything else. I have heard of most of those factors for search, except “Affinity with shared applications”. Not surprising though once you begin reading into the tremendous success Zynga has within their game communities. They target very specific personality types, makes sense that those groups would have shared interests outside the game as well.

  6. I am not really able to get useful results on google. Whenever I search for a topic which is not related to a person, a mix from everything comes up like personal pages, groups, discussions external results but I haven’t found out yet in which particular order, to me it doesn’t seem like the most relevant result comes up.

  7. The majority of FB users don’t have the slightest inclination to read into what they are doing. The wouldn’t even realize that it’s a drop-down menu in the first place. Not *all* users, of course, but, *the majority* are not savvy with technology.

  8. Pingback: » Pandia Search Engine Marketing Wrap-up August 14
  9. Hi Mike,

    You can patent specific approaches to search, and how you might rank content that appears in search results. Google itself was built on the back of Stanford’s PageRank algorithm.

    I think Facebook is failing to take advantage of its search in many different ways, including not making it prominent enough to let people see that they can search for more than just people and groups.

  10. Hi Tom,

    Some of the ranking considerations that Facebook mentions were surprises, such as the shared use of different applications, which might indicate some kind of affinity between the person doing the searching and the authors of content that might be returned, but many of them are things that are listed are things that should definitely be included in a social media strategy.

    I’d love it if Facebook did make it’s search more prominent, and allowed you to do the kind of social search that you suggest, where similarities like geography would be considered (and possibly even something that you could include within your query in some manner).

    This would be an ideal time for Facebook to bring their search more front and center.

  11. Hi Jeremy,

    Facebook’s search isn’t the greatest user experience. Part of the reason why I haven’t been using it to search is because the way it’s set up, it’s easy to jump to the assumption that you can only really search for people or groups.

  12. Hi Keith,

    Thanks. Same here on the affinity with applications as a possible ranking consideration. I guess the popularity of those types of applications are also one of the things distracting us from exploring things like Facebook’s search page.

  13. Hi Andreas

    Useful results on Google, or useful results from Facebook’s search?

    I’m not sure how much of a role relevance actually plays in Facebook’s search. I think they tend to focus upon simple keyword matching more than anything, so for instance, if you search for something like [pizza], you probably would be best served by seeing posts from people nearby writing about local pizza places, but you would probably more likely see results from people whom you are connected to in Facebook first, even if they are on the other side of the country or the other side of the globe.

  14. Hi Mary Lou

    For that reason, I think it makes sense to provide a search that makes it easier for Facebook users to use, and to understand that they can use it to do more than search for other people or groups. The way it’s set up right now, I don’t think that’s obvious enough.

  15. Hi Ben,

    Good points. I think Facebook’s status updates could be used in a manner that’s very similar to twitter, but with the added benefit that people can comment directly upon those status updates.

    I suspect that you’re right that most people who visit Facebook are more interested in staying in touch with their friends, but I think it would be more interesting and useful if they did improve their search interface to make it more obvious that they can search through posts from their friends and from everyone on the site (or at least everyone who is willing to share that content).

  16. I think people generally do not use those Facebook searches and just prefer to use the platform for close social activity from within their friends group. If Facebook were to however have a Twitter-like function to see what people were saying surrounding events that you were going to, I can see people using that.

  17. The only searches I’ve been doing on Facebook are on people’s names. I was unaware that I can actually do “Web searches” until I stumbled upon this post.

  18. Hi Teecup,

    I’m definitely wondering how many people actually do web searches through Facebook because of the way that they seem to hide the ability to do those searches.

Comments are closed.