Facebook introduced an interesting new feature earlier this month called “Graph Search”. In Facebook/Social Media jargon, “graph” is the web of connections formed between you and other people on the social network that are somehow connected to you – friends, family, people who share a particular interest, people who live in a given city or province, and so on. That means Graph Search allows you to search for things – obviously – but differently than you do on Google. Well, sort of.
Uh… What?
Google Search decides how to rank pages based on “relevance”, which is measured with a multitude of factors, including but not limited to:
- Is the page’s content any good?
- Is it really about what it claims to be about?
- Is the information copied from elsewhere?
- How long has this website been around?
Google have also been starting to blend “social signals” into search results – for example links from Facebook and Twitter can help a page rank better because Google assumes people find it noteworthy if they’re talking about it. Integration with Google+ is even tighter; if we’re friends on G+ and I “+1” a page, you will be far more likely to see it rank higher and you’ll see a little “Alessandro +1’ed this page” under it.
So in short, Google do have social factors in their search results, but Facebook search is different. Not better, just different.
Graph Search
Instead of Google’s extremely complex relevance algorithms, Facebook uses a simple factor to decide what it shows you: the Like. You choose your search parameters, e.g. “the restaurant in Moncton that my friends from out-of-town like the most” and Facebook shows you results based on what the people who fit those parameters liked the most. In that example it would look at all my friends that live outside of Moncton and see which local restaurant got the most likes from them.
That’s a pretty good way to get recommendations from friends, something people have been doing since long before the internet existed. There’s a catch though.
Catch-22
Unintentional restaurant pun!
The problem with graph search is in its simplicity. Google Search uses a boatload of factors which allow it to better adapt to less-than-optimal data. If a page has no social signals whatsoever, Google can still tell if that page is worth ranking or not. Facebook Graph Search, on the other hand, is entirely dependent on the Like – it becomes entirely useless if people don’t Like things, or if the businesses in question aren’t on Facebook.
To re-use my restaurant test case, if none of my non-Monctonian friends have Liked any restaurants in Moncton, my search will not yield any results. Furthermore, if I think a restaurant is great but it isn’t on Facebook, I can’t Like it and people will not see it show up in their search. Graph Search only works if there’s a critical mass of business pages and Likes from your friends, which makes this new feature a bit of a “build it and they will come” scenario – it will only become useful if people start using it despite it not being very useful yet.
Graph Search has great potential – and could finally make Facebook more useful for businesses, unlike their abysmally-performing pay-per-click ads – but its success is unclear at this point. People tend to Like a lot of random things on Facebook, so it doesn’t necessarily imply an endorsement… And that’s if they Like anything at all. I, for example, have Liked very few pages, so I would offer almost nothing to Graph Search. Hopefully for Facebook most of its users aren’t like me.