Once UI 2.0 was our strongest launch. It was also the launch that made me question where the software business is heading.We had done the work I always wanted to do for a major release: more complete documentation, a codemod, tests, accessibility and performance audits, and a new Pro scaffolder. The landing page finally showed the quality of the system behind it. To make the foundations useful beyond a demo, I built six apps in parallel. One agent could see the repeated patterns across their repositories and help us decide what belonged in Once UI and what should remain custom business logic.NPM downloads are rising. But visits to our site have been declining, and churn is the highest it has been for us.The tension is hard to ignore: software is becoming easier to produce, while earning a place in someone's life is becoming harder.
The crowded middle
As I built those six apps, I kept seeing versions of the same ideas appear around me. Screen recorders, video editors, mockup generators, social media tools. I was building some of them for my own workflow, and so were other people.That makes sense. If you can spend tokens and a week building a tool that fits your process, another monthly subscription becomes less attractive. The same ability also fills the market with new products, often marketed to the same builders on the same feeds with similar visuals, hooks and offers.When the response is weak, the temptation is familiar: add one more feature, connect one more product, launch again, post more often.I have spent years in that race. Each time I thought I was approaching the finish line, it moved.Distribution is uneven. A good piece of work takes time to make whether it reaches a hundred people or a hundred thousand. Existing audiences compound; people without one have to do the same creative work for far less reach. Social feeds increasingly feel like sales surfaces because so many of us are trying to solve that problem in the same place.For a while, the obvious answer seemed to be: build distribution instead of products.But I have tried the other side of that advice too. Posting more does not automatically create an audience. Launching more can bring more cold offers than real relationships. A viral trial can attract thousands of people interested in credits before they understand why the product exists.I needed a different question. Where is the problem still painful, and where do I already know someone who has it?
Downstairs
A café opened in our building recently. Two friends run it. They made the place beautiful and work there seven days a week.We started talking. I asked whether they had a website. Most of their money had gone into the renovation, they said. An expensive site was out of reach, and neither of them wanted to become a website administrator. They also needed reservations, so a simple landing page would only solve part of the problem.
I know that feeling from the other side. Anything outside your expertise becomes a chain of unfamiliar decisions. Which service? Which plan? Which template? Where do you upload the logo? How does the booking system connect? The work may be easy for a specialist and exhausting for the person who just needs to run a café.
So I proposed a small setup fee and a monthly subscription for a site and a reservation system shaped around their needs. The software could grow from a real operating problem rather than from a guessed feature list.
On paper, building a reservation system for one small customer sounds like a poor deal. Compared with building the same thing for no customer and then buying ads or sending cold messages to find the first one, it looked different to me.
About a week later, I was showing them the site. Another person saw it and asked whether I built websites. She had a small candle shop and needed one.
She looked at the work and said she wanted something like it.
She did not ask whether it used Next.js, Astro, Tailwind or Once UI. She did not care about the animation in the hero. She wanted a site that felt like her shop: restrained, balanced, and focused on the products, with room to choose her typography and colors.
That happens to be where Once UI is strong. The system mattered enormously to how quickly and consistently I could make the work. It did not need to be the thing I sold her.
Since then, three more small businesses have approached me with similar needs: an introduction, a product list, a way to make a reservation. These are early conversations, not proof of a repeatable business. But they are more concrete than another imagined persona in a product brief.
The part a subscription does not solve
There are already tools for almost every one of these jobs. Some businesses use them well. Others run the same workflows through Facebook groups, WhatsApp, email and notebooks.
The setup is a real part of the product. Someone must understand the business, make the choices, connect the pieces and stay around when something changes.
I used to draw a hard line between SaaS and services. I wanted to build software that scaled, not sell my time. Working with these businesses has made that distinction less useful to me.
A small local business may want the software, the configuration, the design and a person who understands what it is for. The monthly subscription pays for an ongoing system; the setup makes it usable. If the same needs keep recurring, the software underneath can become more reusable. The human relationship is what reveals which parts deserve to be standardized.
This does not mean every café will choose the first developer who walks in, or that local work has no competition. Trust, price, maintenance and support still matter. I am learning the shape of those obligations now.
But the first customers did not come from a launch campaign. They came from being nearby, listening, and showing useful work to the next person who happened to need it.
A smaller, more human loop
I am continuing to build Once UI, Aveiro and the rest. The six apps were not wasted work. They made the shared foundation stronger and helped me deliver a working product far faster than I could have before.
What is changing is the direction of the loop.
Instead of building another tool and searching the internet for someone who might want it, I can begin with someone whose work I can see. I can learn how they take reservations, show products or talk to customers. Then I can build the reusable parts and share what I learned openly.
I want my frameworks to remain open. I want students and other builders to see the decisions, the mistakes and the economics, so they can take the same approach in their own communities. And I want the work to leave me more connected to people than another month of trying to win a feed.
The software business is not dying. For me, it is becoming less about persuading strangers to subscribe to another tool and more about making useful systems for people who have a business to run.
The café is downstairs. That is where this chapter started.
