We approached creating Breeze from a single perspective. New, essentially free, phone sensor technology had enabled us to cheaply and easily solve problems that had previously been high friction and expensive ($100+ hardware, remembering to charge, remembering to wear, not damaging in the washing machine).  However, this also meant that the "commoditization of steps" was also about to explode as barriers to entry dropped to zero.  We knew we needed create more value than just "tracking steps" by solving deeper user problems.

We began our process by targeting groups of existing RunKeeper users that we thought seemed likely beneficiaries (older users, folks losing weight, new runners) and simply began talking with them.  Through a good 30 user interviews we learned more about their current attitudes and understanding of activity tracking, health, weight, calorie counting, and steps.  We then used that to create an interaction model and design concept using push notifications to create "awareness" of activity with intra-day engagement loops that kept users learning and interacting without getting bored or overworked.  We determined that awareness (that your actions were being monitored) mattered more than anything else and was the critical path to behavior change, not merely counting steps and analyzing charts.

Or next step was validating these ideas, qualitatively.  To do that we used a storyboard technique, showing possible users a story of how we imaged their problems, our solution, and what we thought the benefit would be.  We had them correct us where appropriate and continued to refine until we were seeing good traction in the lab.  Mind you, up until this point NO real UI was being shown besides sketches to illustrate the concept.

Next we created and interaction model and built out some of the main interface in clickable prototypes:  Onboarding, Realtime views, and Progress views.  We tested these in the lab in parallel to an actual walking-skeleton development prototype, refining the fidelity of the design and the quality of code together.

Finally we had what we believed was a shippable high quality MVP that we shipped to a single test market (New Zealand) where we could continue to refine on-boarding, engagement, and retention metrics until we were comfortable talking about scale.