Recommendations
In this project, I explored the question "How can we help customers find content they will like faster?
In this project, I explored the question "How can we help customers find content they will like faster?
The idea I had in mind was a "Gaming Genome", where the system analyzed the attributes inside games - as more and more games encompass several genres - and recommend games users may not know about. That way, Xbox One could become a trusted place where customers come to discover new content.
However, the first version had to be something we could implement today.
I started by examining Amazon and Pandora as best-in-class examples of recommendation engines that delight customers.
My hypothesis was that customers who like certain types of games will click through to more 'Buy' pages if shown more of the same.
Xbox One, you see me browsing shooters and playing lots of Gears of War, so show me more of what I like.
Then identified the kinds of data we had access to that we could playback to customers:
- Playing/Watching activities
- Rating content
- Friend's activity
- Trusted industry reviews
Then I explored the upvote/downvote interaction for each data type and across several layout specs.
The resulting design updates recommendations on the fly as customers see why they got the recommendation and consequently upvote and downvote content.
