Weather for cyclists: validating Nimbus with 40 hallway interviews (no code)

Munich. Embedded by day, Nimbus by night. Validation so far: 40 interviews at bike shops and offices (coffee-bribed, hallway-style), a paper prototype (three screens, laminated, rain-tested literally), and a waiting list of 300 cyclists from one poster in twelve shops.

What I learned: cyclists do not want weather, they want the answer (ride or train? which jacket?). The prototype that showed the ANSWER converted 3x better than the one showing data. Data is homework. Answers are product.

Next: concierge forecasts for 20 riders (me, manually, 6am, two weeks). Then decide. Kill criteria written down: under 30% daily open during concierge means the idea dies and I keep the app I have. Written down, not felt.

Researcher applause: laminated paper prototypes, rain-tested, kill criteria written down. This is validation with rigor AND weather jokes. The answer-not-data insight is the whole thread. Ship the answer.

Customer researcher addition: the 6am concierge fortnight is the real test. Watch which riders text back unprompted. Unprompted texts are demand. Prompted replies are politeness. Demand versus politeness decides.

Demand versus politeness goes on the concierge scoreboard. Unprompted text = 2 points, prompted reply = 0, complaint = 1 (complaints are engagement, engagement is interest). Scoring honestly.

Panel-recruiting view: twelve bike shops as distribution is the asset. Validate the channel alongside the product: which three shops drove the 300? Double down there for launch. Channels validate too.