Facebook Graph Search: it’s only as good as the data users put in
The other day Facebook excitedly announced “Graph Search” with Mark Zuckerberg stating that “There are so many people on Facebook, you can get a really good signal really quickly.” The problem that we see, however, is that Graph Search is entirely dependent on the inputs that users make. This is the fundamental challenge to Facebook creating a great search product that can challenge Google.
The most important challenge for Facebook is that while 67% of the global internet population of the world have a Facebook account, 25% of those do not use their account at all. When we look at the number of users contributing in an active way, sharing content, photos, comments, links, etc. the number is smaller than might be perceived. The following GWI data from Q4 2012 (fieldwork completed in December 2012) shows exactly why. While 80% have shared photos, only 60% of active users posted a public comment in the past month and crucially, just 72% of those who are active on at least a monthly basis have actively managed their profile in that time.
Growth of active usage is concentrated in passive or frictionless sharing actions and behaviours, like “purchasing a product or service” or watching a TV show or film, which have grown massively through 2012. The other big growth areas are commenting or talking to brands which can be issues we might not always want in the public sphere.
If we look at a market level, the challenge is clearer. The biggest change in users actively posting comments is in emerging growth markets, which are non-English language markets. Secondly, all of the mature Facebook markets are declining as users jump to passive frictionless behaviours.
This is a growing problem for Facebook’s Graph Search, in particular, because active contribution about daily activities is falling in developed English language markets like the US and UK. This type of social data is key to a well-functioning Graph Search platform. The future of social is far more passive, less about interacting with friends and more about watching or utilising the social data to navigate and discover the web as well as the world around us. This means the data will not be user contributed but, aggregated about the user.
This leaves Google with the upper hand. It is far easier to lay social and personalised data over a search product that aggregates the entire internet and links that to users. Facebook is only aggregating what exists in the Facebook eco-system or what consumers have opted into share. To bring this data, which will be coming from Facebook’s partnership with Bing, into the Facebook’s Graph Search will be a far bigger technology challenge and most crucially, an immense privacy challenge.