RoadReview started as a weekend experiment. A friend who teaches driving told us he writes his notes on paper after every lesson and types them up at home. That was 2022. Today the app runs on iOS and Android, in daily use by driving instructors.
Four years later, what started as 320 lines of Swift on a Sunday is a product driving instructors use daily in the car. That's the success story. But in between lie about eighteen wrong turns, a few of which we'll share here.
Wrong turn 1: scaled too early
After the first thirty instructors who loved us, we thought: now we need multi-tenancy, a backend, an admin console. We built for three months. Nobody used it.
Customers didn't want to manage their school, they wanted to teach better. We built for the wrong person. The need for school features came later, and even then differently from what we'd imagined.
Wrong turn 2: treating the data as a teacher rating
Early on we built reports that made instructors comparable to one another. Well meant — for the owners. But the employed instructors shut down immediately: people who feel watched don't use the app. And without the instructors, nothing works at all.
We flipped it: the data is there to help the students, not to rate the teachers. Driving progress is a point of orientation, not an exam grade. Since then, acceptance in the car has been a different story entirely.
What we got right
One thing we got right from day one, and I claim it's the main reason it still runs today: we built with the instructors, not for them.
Concretely: in the first year we invited an instructor every four weeks, paid for a half-day sandwich and let them use the app while we listened. That gave us more than any analytics pipeline.
Don't do that. We never type that. Nobody types that. We have one hour, then the next student comes.Hans-Werner, driving instructor, first test
That quote handed us the main feature: tag buttons right on the map, one click = one mistake. Context is written automatically (position, speed, time). We didn't have to type anything anymore.
What we'd do differently
We should have charged money earlier. Real money. Not in the sense of "the app should be profitable", but in the sense of: people who pay nothing for an app give you no honest feedback either. The first paying customers were the first who really told us what didn't work.
