Evan Bailyn got me thinking last night when he basically said “we don’t necessarily like things our friends like; we like things that like-minded people like.” In other words, just because your friend likes something, doesn’t mean you will. In fact, my experience has been the opposite more often than not.
I think there is a big difference between being friends with someone that your friend is friends with than simply liking the things your friends like (i.e. products, services, art, tv, content, whatever). I think this is because being friends with another human being is an extremely complex relationship. There are literally limitless factors that go into why it is you actually get along. It’s just a more dynamic relationship.
When looking at a product or service or interest, the relationship is more binary and less factors are involved. This means there is a higher likelihood that you won’t align with your friends on everything. A lot of sites and services use this “your friend likes x so you will too” philosophy; Facebook obviously does. But I don’t always care (yes, sometimes I do, it’s not a total blank) — but it’s not nearly as good as showing me what other like-minded people in that particular area like.
Take music for example. There is a good chance I won’t like music my friends like — and having used Spotify and seeing the feed on Facebook, it’s been all too often where I don’t like the band or song my friend is listening too. However, if Spotify were to find other users, who I didn’t necessarily know, that had demonstrated similar interests in music, I think there would be a better correlation to what I like. Show me what they like and come up with a correlation score… sort of like Klout for similarities type thing.
Amazon does this of course, but I think they do it in a rather limited way – you bought this, johnny bought that, you’ll probably like that. I want recommendations to go further – the data exists. I want to know that Johnny bought 50 things in the past that I bought, I feel like if I knew that about Johnny, then the next thing that Johnny bought, I’d probably be pretty damn interested in knowing what it was.
Is anyone doing this btw? For any category (commerce, movies, music, games, articles, whatever)? It would be cool if you could grab this data from likes on Facebook; from my testing, it doesn’t seem that they let you run ads to people that are like minded. I wish they did. Imagine. Show me all of the people that have “liked” these 25 things. These people are like-minded. I want to target them when one of them buys my product. Can you do this on Pinterest? That’s all about likes; they have the data to bucket people with like minds? Can we access this for advertising? Someone chime in.