Launch before you're proud of it.
Ship when the core promise works, not when everything does. A real customer using an imperfect product teaches you more in a week than another month of polishing, and the risk of launching too early is smaller than the certainty of learning too late.
How do I know when the product is ready to launch?
It's ready when one customer can get the core value end to end: they can sign up, do the thing, and pay you. Everything else, settings pages, edge cases, the second feature, can follow.
Reid Hoffman's line has become startup canon for a reason: if you aren't somewhat embarrassed by your first version, you launched too late. The embarrassment is the proof you didn't burn months on things nobody asked for.
What's an MVP for a subscription business?
- 1One customer type, one workflow, one price. Cut everything that serves anyone else.
- 2Charge from day one. A free beta tests politeness; a paid product tests demand.
- 3Handle the missing pieces manually behind the scenes (see: do things that don't scale). Customers care that it works, not how.
- 4Write down what you cut. That's your roadmap, pre-sorted by who complains.
What if early users see bugs and never come back?
Early adopters forgive bugs; they don't forgive broken promises. If the core promise works, a rough edge plus a fast, personal fix builds more loyalty than perfection would have.
What actually loses people: silence. Ship the fix, email them personally, and the bug becomes the story they tell about your support.
Quick answers.
Should I charge during beta?
Yes, even a discounted founding price. Payment is the only user feedback that can't be polite.
How long should building v1 take?
Weeks, not months. On Stencil, days. If the timeline is stretching, the scope is wrong, not the speed.
What features can wait until after launch?
Almost all of them: team seats, integrations, themes, exports, the marketing site's second page. If the core loop doesn't need it, it waits.
Build the thing this guide is about.
Stencil turns your idea into working subscription software — no code required.