Do free users prove a SaaS idea works?

No. Free users prove that people like free things, which you already knew. They tell you nothing about whether anybody will pay, and on a product with running costs, every month you don't find out is a month you pay for.

14 months free

My shared shopping list had more users than anything else I've built. Families, flatmates, couples. People told their friends, which I'd never seen happen before. I used the number of people on it as proof that it was working.

In month 14 I put a small monthly price on it, with plenty of warning. Almost everybody left. One woman wrote to say she'd recommended it to her whole street, and she was right to be cross: I'd let her build a household habit on something I'd always meant to charge for.

What I should have asked

Would anybody pay? That question has a cheap answer. Ask them to, on the first day. The answer arrives in a week. I spent 14 months avoiding it, because the conversation about money frightened me and giving it away felt like generosity.

There was also the category. Households have almost never paid for a shopping list. An evening of looking would have told me.

What free can tell you

How people use the thing. Where they get stuck. What they ask for. That's worth knowing. But usage isn't demand, and demand is the only thing a small product lives on.

The honest version of free

If you want people to try it first, give the trial an end date, tell every user that date on the day they join, and keep it. That turns free into an experiment. A free plan with no end is a habit, and breaking a habit you gave people is how you lose them.

Nobody Bought It has the shopping list's autopsy. Who Pays for the AI? explains why free is even more expensive with a model inside.