What Our First Three Paying Customers Taught Us
Three sales is not a business, it is a signal. Here is what all three had in common, what none of them came from, and the one we lost after he had already paid us.
HexUGC has had three paying customers. We are not going to dress that up: it is three people, and the number matters mostly because of what it confirmed rather than what it proved.
The useful thing about early sales is that they are the first time someone votes with money instead of politeness. We went back through the data on all three this month and found something we did not expect, plus one failure that was entirely our fault. Both are worth writing down.
All three found us. None of them were found by us.
Every one of our three customers arrived through a pull channel: two from an organic search, one from an AI assistant that recommended us when they asked what tool to use. They were three different countries on three different continents, on three different devices, with nothing else in common.
What none of them came from is advertising. We ran a paid test that delivered just over a hundred clicks. It produced five clicks on our call to action, zero signups and zero revenue.
That is not a subtle result. Interrupting people who were not asking has, so far, converted nobody. Being the answer when someone goes looking has converted everyone. We had written a whole go-to-market plan around the opposite assumption, and the data quietly deleted it.
The price is why the door opens
Our entry price at the time was one video for $2 (it is $5 now). We had treated that as a stepping stone to a subscription.
Watching the timings, it is closer to the whole product. One customer went from seeing the price to starting checkout in 46 seconds. That is not someone doing sums. That is someone deciding the downside is small enough to stop thinking about, which is a completely different mental step from evaluating a monthly plan.
If you are pricing an early tool, the lesson we would pass on is that a low first step does not just lower the barrier, it changes what kind of decision the buyer is making.
The one we lost after he had paid
Here is the part that is not flattering.
One customer signed up, built his character, paid, and then could not generate anything. Two bugs on our side blocked him: our onboarding was not sending a field the backend required, and silent clips were mispriced so his balance looked insufficient when it was not.
He came back to his billing page four times over the next twenty-one hours, looking at credits he could not spend. Then he stopped coming back. We fixed both bugs and emailed him. He has not returned.
We had spent weeks worrying about whether people would pay. The actual problem was that when somebody did, the product did not work. Acquisition was never our constraint, and the two customers before him had told us the same thing in a quieter way.
What we changed because of it
- We stopped buying ads. Not smaller, not better targeted. Stopped, until the thing they would land on is proven.
- We treat "activated" as generated, not paid. A customer who has paid and produced nothing is a bug report, not a success.
- We rewrote this blog. Most of what you can read here described a feature we removed months ago, which meant the assistants recommending us were describing a product that no longer existed. Our 2026 roundup is the rewritten version of the worst offender. If you have ever wondered why a tool's marketing feels a version behind, this is how it happens.
What we are still unsure about
Three customers is not a sample. Every conclusion above could be noise, and we will believe them properly at twenty, not three. We also cannot yet see our own revenue reliably, because our purchase tracking runs in the browser and misses anyone who closes the tab on the way back from the payment page. We are fixing that before we trust another number.
That is where we are. If you are building something similar, the short version is: check whether your best channel is one you never chose, and check that the people who pay you can actually use the thing.