According to a recent article in Advertising Age, Google’s social strategy does not involve building social networks to compete with Facebook but rather it is focusing on using social data to build better applications:
“As an example of the current strategy, [Eric] Schmidt talked about getting more information from YouTube users in order to offer more targeted video.”
YouTube already has a fairly robust recommendation engine:
but from Schmidt’s comments, development around this area of exposing social analytics is where they see real business value. This is backed up by moves such as YouTube’s purchase of movie recommendation site Fflick.
How can analytics be used to derive value?
For instance, predictive analytics solutions (like IBM SPSS) can traverse a large inventory of content and make associations based on a visitor’s past behavior and the behavior of their friends in the network. Match this with sentiment analysis, which can be used to look at the conversation around a video to determine whether it is loved or loathed (or somewhere in-between), and suddenly you have a more immersive viewing experience.
This doesn’t just apply to Google and video. Foursquare is apparently taking this approach to differentiate itself as Facebook encroaches into its space with its Places offering.
Whilst analytics can offer differentiation in a hotly-contested area, the issue of privacy has to be addressed. The interfaces can get so good at offering recommendations that they border on being plain creepy. Couple this with the growing paranoia around the extent to which our digital lives are tracked, and suddenly these interfaces appear more Big Brother rather than benevolent Jeeves. One way to address this issue is to be as transparent as possible when exposing social analytics.
So if Eric Schmidt’s comment can be taken at face value, I’d suggest it’s in the context of a growing trend in looking to maximize the value in existing networks rather than racing to build new ones. Social analytics, when handled deftly, can unlock this latent value in social data.
Do you agree?