It seems simple when you hear it, but one of the most interesting examples of social media data I have read about recently comes from Troy Carter, Lady Gaga’s manager, at last week’s Wired 2012 event. Using social data from Spotify to help to design the set-list for a gig. Simple, effective and the kind of personalisation that is only available with widespread social sharing.

As Carter puts it:

I can sit down with the guy from Spotify, and he shows me this spike on Fridays as people listen to Gaga before going to the clubs. When I go to South Africa I know to include this song in this set, because I know that’s a fan favourite, and also to take this song out. We’ve never had a direct relationship with an audience. When someone buys a CD we used to count them as a fan, but we never knew if they hated the CD and threw it out the window.

So simple it’s obvious – look at which tracks are popular and which are less popular in the area around a gig and then play back to the audience what they want to hear. You could then go a step further to introduce new tracks based on these data patterns – introduce new content into the set based on what people already enjoy listening to, and leave out new tracks they might be less keen on.

This kind of personalisation is clever and could be used much more. It is not based on an individual’s social data but on looking for trends and patterns among a group of the population.

Looking at the data people leave behind from the actions they do in social media (what they say, do, listen to, watch) presents a huge opportunity for many brands and products. And if you layer in data that only you have (eg purchase history) you can produce an even more powerful data set.

Brands need to start understanding the data they have, and the data they can learn about their customers through social media. Only then can we start to be more creative about how we are using this data. And only then will we see more examples of something as simple, and as sensible as how Lady Gaga personalises her gigs.

Original Post: