More effective "Interestingness"

Tim O’Reilly wrote about Interestingness today; a lot of people have written about it.

For those of you that might not be familiar with the term, it is basically a measure of value someone places on something. That something could be a photo on Flickr, a blog entry, some gadget for sale, an eBay auction – basically anything that can be rated.

As Tim says,

Whether it’s pagerank at Google, interestingness at Flickr, or diggs, or SmugMug’s most popular feature, we see all across the web attempts to incorporate human intelligence into web applications. As I’ve written many times, harnessing collective intelligence is the very heart of Web 2.0. And that intelligence is distinguished by its bionic nature: we’re building applications that are a fusion of human and machine.

So basically, there are numerous ways people tell us they are “interested” in something. From voting (like MSNBC’s The Week in Pictures) to clicking on a “Thumbs up, Thumbs down” type rating system.

These are useful ways for us to find out what the aggregate population thinks of something. And that’s fine – the data might be valuable, and it might be interesting – but knowing what EVERYONE thinks about something is not as valuable to me as knowing what other people that are “like me” are interested in. And this is where Interestingness – and the plethora of ways of measuring it, displaying it, etc fall short – they capture the fact that x people thought something was interesting, and that y people didn’t. But they are missing the opportunity to capture more – such as of the x people, 65% of them also read the Dilbert cartoon and maybe 43% of them also own an XBOX360.

Whatever the other data points are, the accumulated data would allow me to not only know that most people find an item interesting, but the data would allow me to find out how many people like me thought it was interesting – and perhaps even more important – what other things do people like me find interesting?