Amazon, it’s a jungle out there. Photo: Michael Durwin

One of the important aspects of Anderson’s Long Tail and the Unsymposium this week was the interconnectivity of the network and the recommendation engines of online shopping sites and services. As I touched on last week, I’m a fan of services like Spotify and other ways of buying “things” online. One of my reasons for liking Spotify is the artist recommendations that I get, based on what I listen too. It opens up worlds of exploration. However as was discussed in the Unsymposium, there are some issues regarding these systems. They are mainly based on the technical component of the recommendation; meaning the engine that runs the system as Adrian said. Services like Facebook make recommendations that are based on advertising and money making. They should be ignored, or at least dealt with more thoroughly. Something like Amazon or Spotify make recommendations based on what you and other people are interested in, and how one product relates to another. This is a much better system and one that has much more potential.

I have a bone to pick with these systems though. When I first started to buy things from Amazon, I thought great! Look at all these interesting books, movies and albums that I would have never thought to look at. Now I’ve been buying things online for years and they’ve harvested a lot of my personal data. Unfortunately I have varied interests, like most people. Sometimes I get a good recommendation. The rest of the time, I don’t. I have sent Amazon’s emails to my spam folder. The problem? As soon as I show interest in something that is incredibly popular, say for example a blu-ray like Marvel’s The Avengers, my personal recommendations become over powered by garbage. Millions of people like that movie, they look at it, wish list it and buy it. That means I start getting lots of recommendations for the big box office hits of the year. The problem with this? Unlike a lot of people who buy movies, I’m not really interested in the biggest and best movies for the most part. What began as a personal recommendation has been polluted and violated by trends. I guess that’s the problem of buying a huge range of different items from these sorts of services, it becomes almost like a grocery store. It wouldn’t make sense for Woolworth’s to recommend carrots to people who buy bread, simply because so many people do large grocery runs that contain both. Sure some people might be interested in both, but if you’re looking for a recommendation, it needs to be more specific. Perhaps these systems in their current form are best for purists and people who tend to purchase only products from a niche market online. I guess their potential is limited by the current software and engines that run these systems.

So in some ways, there are issues with these systems limiting the scope of what you buy, like Elliot said. But I’d also like to believe what Jasmine said, about it not being an issue because people are capable of ignoring the system. I think media literacy comes into it again. As someone who has grown up with these things coming into play and has studied media and advertising, I’m pretty switched on when it comes to these recommendations. I’m confident enough on the web to make my own decisions, and know when a recommendation is relevant or simply just another popular pick. For me, they don’t really create an issue, other than the fact that I don’t get to reap the benefits of them properly. A lot of people who aren’t as experienced with web based media and recommendation systems, don’t take recommendations lightly, they’ll follow up on them, especially if they’re something super popular. What it really comes down to is your literacy and knowledge of your own tastes and interests, in combination with recommendations. On Spotify I tend to ignore a lot of recommendations, because they’re bands I don’t like. There’s a few bands I like that are like a diamond in the rough. Maybe one band out of a genre that I find entertaining. Where Spotify assumes that I like thinks simply based on what other people like and similar music, that’s not really how my tastes work. Perhaps I ask too much of the system? I think though that they are definitely a positive factor of the network and something that I can see being developed and implemented to become more personal and effective in the future, as Adrian said.