Geekularity

Monday night, Sean Kane, the Director of User Interface Engineering at Netflix, gave his presentation entitled “Developing a Great User Interface” to members of the East Bay IT Group . I found it particularly fascinating to see how Netflix integrates their feature development and testing into a very rapid, biweekly release cycle. Most of the features, once they graduate to production grade code, will exist on the production web site and will only be available to one or more select groups of users; usually groups of about 10,000 users each. Depending on the feature and what usability metric they are testing for, these test groups will last for several weeks or several months. Other test groups might run concurrently and will often overlap. This testing system sounds fascinating, especially when I realized that I may have been a member of one of the test groups in the past!

One point I found especially cool was Sean’s discussion about the red star ratings system. Many Netflix users, including myself, believe that the red stars are the average rating from everyone who rates that movie. It turns out that this is only true when you first subscribe to the Netflix service and your ratings and movie queue are nil. Once you have added movies into your queue or begin to rate other movies, the red star rating becomes personalized to you. Based on you viewing, rating, and queuing history for example:

A “Weekend at Bernie’s” search while not logged in and my cookies cleared

Same search, but this time I’m logged into my account

Apparently Netflix thinks I am not as much of a fan of “Weekend at Bernie’s 2″ as the rest of the Netflix population. To the contrary, I loved it! I’m a sucker for both of those movies. Just after I took those screen shots, I rated them both 5 stars!

But this isn’t about my taste in super awesome movies, this is about user experience. As Kane even admits, Netflix needs to do a better job explaining some of the finer points of their user interface. Perhaps a delayed pop-up bubble when hovering over the ratings object could present the user with a “did you know” type explanation with a link to a details page that explains the ratings features in depth. While this feature is optional and would not divert too many users away from the task of hunting down movies to watch, it would give the geekier of us the chance to understand the system a bit more.