Tag Recommendations

This time it’s personal

Web 2.0 was about making the web social (and glassy buttons), now let’s make it personal, relevant and about the user.

There’s so much content it’s hard to filter out the noise and get to the stuff you want. This is true be it from Twitter, your Facebook news feed or what to watch on TV. The growth of smartphones only exacerbates the need for a personalized experience. Our phones have become an extension of ourselves, though their smaller form factor requires us to only put the important stuff on them. The way we use our phones also dictates a need for faster access to the important things. Besides streamlining the features and the design, the mass of content needs to be streamlined as well. Quicker access to the things that matter to you is the core concept behind Microsoft’s current ad campaign for WindowPhone7.

The mobile space isn’t the only place where this streamlining is welcomed, take Netflix for example. They’ve grown from a simple DVD-by-mail service to one of the biggest online streaming services. There’s a reason people love Netflix. It’s not about the number of movies they have but rather they showcase the videos that you may actually want to watch. When first signing up to the service you’re asked to rate a few movies so it can begin to make recommendations. Netflix even had a ongoing contest looking for anyone that could significantly improve (their already lauded) recommendation algorithms. In Sept. 2009, they had a winner but the real winner was Netflix and their customers.


Personalization doesn’t always need to be complicated, even the smallest touches of personalization will do wonders for the user experience. The latest browsers have removed their default homepages in lieu of quick views of sites that you visit most. Above is the message Safari displays when you launch it for the first time. Like Netflix, it’s aim is to show you content that’s relevant to you based on your actions. Below is a basic recommendation system I created to demonstrate how user actions can be used to bias content towards more items of a similar nature (in this case color).