Amazon Fire TV Cube
Amazon Fire TV Cube is the first product to combine far field voice control with an on-board universal remote. The original vision sounded quite simple: Customer should be able to walk into the room, say “Alexa, watch NBC” and Fire TV cube will turn on your TV and make sure NBC is playing.
Fire TV Cube is built on the Fire TV UI that I worked on 2 years back, with 2 additional key features: far field voice control and universal remote capability. In this project, I was the lead designer on the universal remote. There were 2 big issues that I was trying to tackle:
- Universal remote is based on old IR technology, but far field voice assistant is still quite new as I was writing this in 2018. As voice forward experience is all about minimizing friction, IR technology is notoriously known for generating friction.
- The bulk of the work was in set up. We know from extensive established Fire TV research that customer want to get through setup as quickly as possible, but a first glance of the process did not seem uplifting: it was almost twice as long as the overall flow that it was supposed to go into.
Neither me or my PM partner wanted to triple the set up flow. Tons of work went into optimizing the flow so that minimal amount of customer input is required to teach Alexa how to control TV and AV receivers. Since my team is situated inside a larger internal design studio at Amazon, I had great resources and design partners to help with the work. Our process at a high level looks a bit like this, from my perspective:
The bulk of the work was on me and the UX writer. It was my job to decipher technical and business requirements and turn them into customer facing designs; with help from the UX writer, we presented the designs in ways that regular people can understand. We then worked with user researchers to set up lab studies with real customers; this was where we tested the designs and iterated based on real customer feedback. Although I was not the one interfacing directly with customers, I got to observe the study and received feedbacks in real time.
However, with this particular product, given its technical complexity, there was another step before all the standard design process happened: how do we avoid tripling the set up process? The final answer came from engineering innovations and lots of agonizing hours of flow optimization. How can we move things around to eliminate one screen? and one more, and maybe one more?
My sketch files looked a lot like those crazy mind maps you see on TV, when a detective (professional or self-self proclaimed) is trying to map out clues to solve a case. I looked at every single set up flow from a high level to make sure they were as frustration free as possible. The task forced me to back out of visual design decisions so that I can focus on the overall experience. It was only when the flow was in the best shape it could be, then I start working with UX writers and visual designers to fine tune each screen to it’s perfection.
After all, we shouldn’t make customers work for something as simple as turning on their TV.