Skip to main content

What's good (and bad) about Facebook graph search

On Tuesday, Facebook announced it will be adding search functionality to the social network. The new Facebook graph search feature and functionality will roll out ever so slowly to a small set of users while the company continues to test and fine tune it. You can sign up to be on the waiting list of beta testers now.

That's all good, but let's take a closer look at what's potentially good and bad about this new feature.

Analysts, writers, and Facebook users have long wanted a search tool that would find more than just Facebook accounts – people, businesses, and fan pages. They wanted a simple way to find a link to an article that one of their friends shared, when exactly? A week ago? A month ago? Who remembers these things? Those of us who use the Internet and technology rabidly typically don't worry about remembering because we rely on search functionality. Search tools are the dashboard to our second brain. Facebook hasn't had it before, and gosh darn it, isn't it time that it did?

Meanwhile, you can look for not only posts, but also images. Double win.

But those first two pieces are only a small part of the puzzle. Judging by the demos and announcement content, it seems Facebook really wants us to know how else we can use search to find out interesting information about people in our network and what they like.

A mix of good and bad

This is the point – when specific use cases were touted – when I started to get apprehensive about the yet-to-be-released features. You can find out what you want to know about your friends using natural language searches, such as, "Friends who like Star Wars and Harry Potter," or more to the point, "Cameras my friends like" so you can potentially make a purchasing decision based on information from your friends.

First of all, I don't want someone else to tell me how a search tool should work or what I should be using it to find, and I got the sense that Zuckerberg and company really wanted to impress upon us Facebook users the way they envisioned graph search working.

I didn't like feeling as though I was being ordered to use it in a certain way, a way that will help Facebook monetise its search features. Granted, it's helpful to know what a tool can do, even if that's not what many people want to do with it. Better still, the official announcement was full of refrains about how Facebook would thoroughly test the new search tools and make changes based on usage. All told, that's a mix of bad and good.

Clearly, though, this is the angle from which Facebook graph search will be monetised. Many of the examples cited in the demo seemed to set Facebook up to compete with other recommendation sites, the likes of Yelp for example, and even online dating services.

Facebook's search tools are not, in my mind, competing with Google in any way, shape, or form. Google's search primarily (although not exclusively by any means, and certainly not since Google connected all a user's Google accounts to inform search results, although that's another matter) looks at nearly everything on the Internet. Facebook and other recommendation sites only search their own networks and rely on users to contribute to the pool of information they maintain – the database. No one uses the term "user-generated content" much any more, but it's exclusively what powers Facebook from a content perspective.

The real problem with Facebook graph search comes down to the implicit assumptions Facebook is making about what must happen in order for the search to work and the results to be relevant.


  • The place, product, brand, doctor, service, or whatever it is you hope to find, exists as a Facebook business or brand page.
  • That page must be promoted or somehow discovered by people in your network.
  • Your friends must have truthfully "liked" that business/brand page on Facebook.
  • You trust your friends' opinions and recommendations.

And let's dive into that third point just a little bit deeper.

What My Friends Actually "Like":

  • A friend who works in the non-profit field happens to "like" a lot of non-profit organisations, causes, and policy advocates.
  • A male friend who works in advertising "likes" some feminine hygiene products, a few no-calorie sweeteners, and several brands of women's beauty care products because he works on those accounts.
  • My mother "likes" the "I (Heart) My Daughter" community (aw), a church she hasn't been to in a year, and the community "I don't care how old I am, I WANNA GO ON THE BOUNCY CASTLE!!" (sic).
  • One of my sisters "likes" just about every yoga studio within a 50-mile radius of her house. She also "likes" the movie Nacho Libre, which I find adorable given her other interests.

The point here is what we "like" on Facebook is often not a clear and accurate representation of what we actually like. We "like" things because in the past, the concept of "liking" something on Facebook was for fun. (Too many Facebook users never understood that "liking" a fan page or business was feeding other business metrics, but that's beside the point for now). And now, Facebook is changing the rules of what it means to "like" something.

We've heard that before: Facebook changing the rules after its users have already participated in the matter at hand.

I'm much less concerned about privacy issues in this instance. The new search feature claims it will only make information available to you if it's already been shared with you. And if you or your friends go around "liking" products and services, or telling Facebook where you live, and so forth – that's on you.

If you care about privacy, you should remove that information from your profile and remove "likes" that might give the wrong impression about your identity. And be sure to do that now and over the next few weeks before the search feature rolls out in full.