Recommender Systems

I was really interested in this week's readings on recommender systems and how they can affect users' opinions on specific products that a service has to offer. When using applications such as Netflix or Pandora, they will recommend movies, TV shows, or music that you may like based on your previous selections. These suggestions don't force you to choose them, they are simply suggestions, but they can help open a user's eyes to new content they have never heard of before. These recommender systems are formulated by specific algorithms in order to determine what each specific user may like. If I like watching comedy movies on Netflix, then the "suggested" section of my Netflix account will mostly be filled with other Comedy movies that I haven't yet watched on Netflix. Pandora's entire experience is based on recommendations and suggestions. When you choose an artist, or a band, or a specific song that you like, it creates a whole personalized radio station around that one choice and gives you a playlist of other songs related to that one.

These recommender systems aren't perfect. Maybe you don't like some of the suggestions that they give you. This is where Pandora's "skip" button comes in handy. Maybe you share a Netflix account (because almost everybody does) and the suggested movies and television shows are programs you would never even consider watching. There are a few flaws in these algorithms, but the majority of the time these recommender systems really help, and they've definitely been the reason behind some of my favorite features to watch on Netflix.

Some questions do arise from this week's information, however. As I just previously mentioned, there are some flaws with these recommender systems. What can companies like Netflix do to improve their "recommended for you" section of their application when dealing with shared accounts? Can Pandora recognize the songs you skip and go even further to make sure songs by that particular artist don't show up again on your playlist? How influential are the "thumbs up" and "thumbs down" buttons on Pandora? Do many users take advantage of the recommender systems offered by these companies, or are they something people simply ignore? Do ratings by other users really affect how many others interact with that specific product? These are just some of the questions that I thought of while reading through the recommender systems article for this week.

-Frank Hines