Before you start building your own station smart speaker skill, “Think, Make, Check”

Feb 20, 2018

Voice assistant platforms like Amazon Alexa and Google Assistant are currently experiencing an adoption rate that outpaces that of smartphones in the early 2000s. As our Chief Digital Officer Tom Hjelm outlined in a note to stations, these platforms represent “a huge part of our shared future, and when you ask Alexa, Siri or Google to play your local Member station, or NPR, or the headlines, or your favorite podcast or show, we want to ensure the experience is as direct, relevant and compelling as possible.”

So, how does a team try to make that vision a reality when deciding what to build?

Likewise, during our Smart Speakers 101 webinar, a station asked, “Should we build our own Alexa skills?”

A challenging question to answer unless you break it down. There are many things to consider and explore before you embark on development: what are your resources, can you bring unique value to your audience, how will you bring awareness and promote your product, and how can you create continued engagement. For anyone facing a question like this about building a skill — or other product features — you can get to the bottom of what to build by first identifying why you would build it and who you would build for.

I want to share some insight into the approach that our team took as we started thinking about voice assistance. We were trying to get to the bottom of how to make station streaming a reality on Alexa. For that new NPR station streaming skill, we applied the tools of design thinking to the Lean UX model of Think, Make, and Check. This model is rooted in a few foundational principles:

  • The team shares an understanding of what success looks like.
  • The thing to be built solves a distinct problem for real people.
  • The team employs continuous validation at every step, internal or external.

Here's an overview of the process we followed.


Our first step was to ask ourselves how might we improve listening to NPR live streams on Alexa so that users are more successful. The team worked to define the problem statement, generate assumptions on how to fix the problem, and then prioritized what to fix first.


Next, we created prototypes, validated our assumptions internally, and tested our solutions externally. Throughout this development phase, we continued to ask if the prioritized features represented the most effective ways to meet our goals: 

  • More awareness of station streaming
  • Better accuracy of station names
  • More transparency around what stream is playing
  • Better wayfinding within the streaming experience

We prototyped the voice flows in Google Slides and iterated on them internally. Once we felt confident enough to test with real users, we fielded a very lean user test.


Our lean user test yielded valuable feedback that the team absorbed, and used to iterate on the prototype. This phase allowed the team to learn more about our users and the platform itself; this better understanding led us to shorten the onboarding flow and change our views on the value of the information we display within the Alexa app.

Voice assistant platforms are evolving quickly, and user behaviors are evolving and adapting to this new user experience just as rapidly. This fast evolution provides both challenges and opportunities: sometimes we hit a wall of “No” on what felt like basic functionality that we wanted to build, while at other times those restrictions were suddenly lifted by the time we revisited the prototype to iterate.

Our biggest takeaway from this project is the value of applying the Lean model, and to move through the phases as efficiently as possible, so that you can get to and validate the highest priority user problem before you start to build.

To sum up, to get to the bottom of what to build, first identify why a custom skill and who is it for. The "why" for stations is - why is their custom skill valuable and the who is their local audience. Don't forget to research how your station resources will be impacted, especially as you consider ongoing support, promotion, and engagement for your skill. Read more about our answer to this question on our Smart Speakers FAQ

Stay tuned for more blog posts breaking out some tips and tools for each of the phases.