The shape of small SaaS

When I started selling software, the barrier felt enormous. You needed to know how to code, how to deploy, how to handle payments, how to manage servers. The weight of all that kept most people out.

Now the landscape looks different. No-code platforms have matured to the point where they handle the heavy lifting. And AI tools like Claude Code have changed what one person can build alone.

I built a small internal tool last year using a no-code database connected to a simple frontend. The entire thing took three evenings. The infrastructure costs are twenty dollars a month. The tool now processes around four hundred requests a week for a niche audience that pays a flat monthly fee.

What made it possible was not raw technical skill. It was patience. I did not try to build something massive. I built the smallest possible version of something that solved a real problem I understood deeply.

People assume the hard part of passive income is the building. I have found it is something else entirely.

The hard part is staying with the problem long enough to understand it. Code is just the expression of that understanding. If you build something that solves a problem you yourself have, you skip the guesswork. You already know what matters because you live inside the gap every day.

I talked to someone recently who built a small SaaS for writers who struggle with formatting submissions to literary magazines. He was one of those writers. He knew exactly what confused him because he had been confused for years. The tool took him two months to build with no-code tools. It now brings in enough to cover his rent.

He told me the building was the easy part. The years of being frustrated with the problem were the real investment.

I have settled into a rhythm that feels sustainable. It looks something like this.

I start with a problem I have personally. Not a market gap I researched, but something that genuinely irritates me in my own work. I write down what the solution would look like if it took no effort at all. Then I build the absolute minimum version using a no-code tool like Glide or Softr paired with a database like Airtable.

For more complex logic, I bring in Claude Code. I describe what I want in plain language, let it generate the scaffolding, and then I adjust. The AI handles the parts I do not know how to do myself. I handle the parts that require judgment about whether the solution actually feels right.

Then I put a Stripe link on it. Not a free tier, not a trial period, just a simple monthly price. I want to know immediately if someone will pay. If they will not, I want to know that before I spend months adding features nobody asked for.

Pricing used to feel complicated. I spent hours looking at competitor pages, trying to match their tiers and their language.

I stopped doing that. Now I pick a number that feels slightly uncomfortable but not outrageous. Usually that lands between nineteen and forty-nine dollars a month for something that saves someone a few hours of work each month.

The pattern I have noticed is that people do not care about the price as much as they care about clarity. They want to know what they are paying for and they want to know you are not going to change it on them every few months.

I have a friend who runs a small tool for podcasters. He charges a flat thirty dollars a month for unlimited use. No tiers, no add ons. His customers stay for years because they never have to think about what plan they are on.

Claude Code has been a quiet shift in what I can attempt. I do not use it to generate entire applications. I use it to generate the parts I find tedious. Forms, database queries, the glue between pieces.

What I have noticed is that the AI is good at syntax but bad at context. It does not know why someone would want the thing it is building. That part still belongs to me. I spend more time describing the problem clearly than I do debugging the code it produces.

One workflow I have repeated a few times now: I map out the logic in plain English first. I write a document that describes what the user does, what the system does in response, and what happens when things go wrong. Then I hand that document to Claude Code and ask it to generate the pieces I need.

The result is not perfect. I still have to test everything, adjust things, make sure it actually feels good to use. But the time from idea to working prototype has gone from weeks to days.

The passive income fantasy usually ends at the part where you have to support what you built. Customers email with questions. Something breaks on a Sunday. A feature request piles up.

I have learned to design for that. I keep the surface area small. One integration, not five. One clear use case, not a platform. The less there is, the less there is to maintain.

I also price knowing that support takes time. My monthly fee is not just for the software. It is for my attention when something goes wrong. If the price is too low, I resent the time it takes. If it is too high, I feel pressure to add complexity I do not want. The right price is the one where I do not mind answering an email on a Tuesday afternoon.

Optimizing for growth? Go-to-Millions is Ari Murray’s ecommerce newsletter packed with proven tactics, creative that converts, and real operator insights—from product strategy to paid media. No mushy strategy. Just what’s working. Subscribe free for weekly ideas that drive revenue.

Sign up now

I have noticed a lot of excitement around AI tools that promise to build entire businesses automatically. That excitement feels misplaced to me.

What AI actually does is remove the friction of implementation. It does not remove the need for taste, judgment, or patience. If anything, it makes those things more valuable. When anyone can build something quickly, the only differentiator is whether you built the right thing.

I use Claude Code because it lets me skip the parts I do not enjoy. I still have to know what I am trying to build and why. That part has not changed.

The products that have lasted for me are not the ones I built fastest. They are the ones I kept returning to over time, adding a little here, simplifying a little there.

Passive income is not passive in the sense of doing nothing. It is passive in the sense that the work compounds. You build something once and it continues to serve people without requiring your constant attention. But that only works if you built something worth serving people in the first place.

I have a small SaaS that makes about eight hundred dollars a month now. I spent maybe ten hours on it last year. Those ten hours were mostly answering questions from users and tweaking the onboarding copy. The year before that I spent forty hours on it. The year before that I spent the initial burst of building.

That is the shape of it. A concentrated period of focus, then a long tail of small refinements. Not glamorous. But quiet in a way I have come to appreciate.

I do not think everyone needs to build a SaaS. But I do think there is something valuable in learning how to build a small system that runs without you. It changes how you see your own time and what you are capable of.

If you have been sitting on an idea, the only thing I have noticed is that the people who eventually build something are the ones who start small and stay patient. The tools are there. The rest is just showing up to the problem long enough to understand it.

Next week I want to write about something I have been thinking about for a while. The quiet tension between consistency and novelty when you are trying to make something that lasts.

AI promises speed and efficiency, but it’s leaving many leaders feeling more overwhelmed than ever.

The real problem isn’t technology.

It’s the pressure to do more with less — without losing what makes your leadership effective.

BELAY created the free resource 5 Traits AI Can’t Replace & Why They Matter More Than Ever to help leaders pinpoint where AI can help and where human judgment is still essential.

At BELAY, we help leaders accomplish more by matching them with top-tier, U.S.-based Executive Assistants who bring the discernment, foresight, and relational intelligence that AI can’t replicate.

That way, you can focus on vision. Not systems.

Download the 5 Traits AI Can’t Replace.