What shipping my first paid app taught me about pricing

I spent three months building FlickClean and about twenty minutes deciding what to charge for it. In hindsight that ratio was completely backwards, and almost everything I got wrong in the first month traces back to it.

Here is what I would tell myself before that launch.

Nobody pays for something they have not felt working#

My first build showed the paywall on launch. It converted at [[under 1%]].

The reasoning seemed sound at the time: people who are not willing to pay are not worth onboarding. What actually happened is that nobody had any idea what they were being asked to pay for. A photo cleaner is not a category people have strong intuitions about.

Moving the paywall to after the first completed cleaning session — once someone had watched a few hundred photos disappear and felt the relief of it — changed conversion by roughly [[3×]]. The app itself did not change at all.

Price is a signal, not just a number#

I launched at [[CHF 1.90]] a month because it felt modest and I was nervous.

Raising it to [[CHF 3.90]] barely moved conversion. What it did move was refund requests, which went down. My read is that the very low price made people suspect the app was thin, and they bought it with low expectations that were easy to disappoint.

The cheapest version of your product is not the easiest one to sell. It is the one people trust least.

Annual plans carry the business#

Roughly [[70%]] of revenue comes from annual subscriptions bought by people who had been on monthly for less than [[two weeks]]. Once someone has decided the tool is useful, they would rather stop thinking about it.

If I had launched monthly-only, I would have concluded the app was not viable.

Instrument the funnel before you need it#

For the first three weeks I had no analytics beyond App Store Connect, which told me how many people subscribed and nothing about the ones who did not.

I was making pricing decisions from a total of [[four data points]] and a feeling. Every real improvement came after I could see where people stopped.

Concretely, the events worth having from day one:

• App opened, first time versus returning
• Permission granted or denied
• First cleaning session started, and completed
• Paywall shown, dismissed, purchased
• Subscription cancelled, with how long they stayed

That is five minutes of work before launch and the only reason I could tell the paywall move had worked rather than guessing.

The part that actually mattered#

None of the above is really about pricing. It is about the fact that I built for three months without ever watching someone use the thing.

The twenty minutes I spent on price were a symptom. I had treated shipping as the finish line, when it was the first point at which I could learn anything real.

The next project started with the funnel and worked backwards, and it was a better project for it.