From idea to product: lessons from building SaaS

I have built products the way most builders do: one big idea, too many features, a long road to the first real user. Building Wichtelkompass and eventOz taught me the opposite order. These are the notes I wish I had when I started.

Scope is a daily decision

Every product starts with a vision that is three products wide. The discipline is not deciding the vision once; it is re-deciding what not to build every single week. A childcare platform that tries to serve families, providers and institutions equally well on day one serves nobody. Pick the thinnest slice that creates real value, and say no to everything adjacent until that slice works.

Ship to reality, not to the roadmap

Roadmaps describe intentions; users describe reality. The gap between them is where products die quietly. Shipping early (even unfinished, even embarrassing) replaces assumption with signal. The features users actually need are almost never the ones in the original plan.

Boring technology is an advantage

For a small team (often a team of one), every technology choice has a maintenance price. The stack that lets you ship this month matters more than the stack that scales to ten million users next year. Modern, boring, well-documented tools (Next.js, Postgres, TypeScript) buy speed now and optionality later.

Sustainability beats scale

The overnight success is a survivorship artifact. Products that last are built by people who can keep building: sustainable pace, sustainable costs, sustainable motivation. That is a reason to run lean infrastructure, automate the operations you can automate, and design the business so it funds the next twelve months, not the next twelve weeks.

What I would tell myself at the start

1. Talk to the people who will operate your product daily: they know where the value is.
2. One working feature beats five almost-working ones.
3. Distribution starts before the product is done, not after.
4. Write down what you learn, or you will pay for the same lesson twice.