When to turn your custom app into a product

You built an app that runs your business. Here is how to tell if others would use it, and the five things that change first.

Say you built an app for a flower shop. It tracks the money, plans the delivery routes, messages customers when their order leaves and takes payments. The database is connected and the shop runs on it every day.

Then another shop owner sees it and asks if they can have it too. That question starts a different project from the one you finished.

Is your app worth turning into a product?

It is when other businesses have the same problem and solve it the same clumsy way. One request is a signal. Three businesses you could name today that work the way yours does is worth testing.

It is not when the app is mostly one owner's habits written down. If every screen exists because of how one person likes to work, the next business will want half of it changed, and you will be building a second custom app rather than selling the first one.

More people are building their own software with AI tools, and many of those apps work well for the business they were made for. Far fewer are ready for a second. The gap between the two is architecture, and it comes down to five changes.

Does every record know which business it belongs to?

Not yet, if the app was built for one business. Every order, customer and invoice in the database is simply yours. In a product, each record carries the business it belongs to, and the app checks that on every request, so one shop can never see another shop's customers. This is the change that is hardest to add late, so it comes first.

Is your own business written into the code?

Almost certainly. The shop name, the delivery zones, the opening hours and the way a discount works all tend to sit in the code. A product keeps one shared core that every business uses, and moves the parts that differ into settings or modules a business can switch on. One update then reaches every customer, instead of one copy of the app per business.

Can each business bring its own payment and accounting tools?

Only if the connections can be swapped. Your app talks to one payment provider, one accounting package and one way of sending messages, hardwired. The next business uses different ones. A product treats each of these as a connector: the app asks for a payment, and the business decides which provider handles it.

Can a new business get started without you?

It has to, or every new customer is a small project. That means a sign-up that creates a new, empty business, roles so the owner and the staff see different things, and an import for the customers and orders a business already has somewhere else.

Who keeps it running once other businesses depend on it?

Someone has to, every day. A product needs billing, backups, monitoring that notices a problem before a customer does, and a security check. Apps built quickly with AI tools often leave keys in the code or skip the check that a user may see a record at all. With one business that is a risk. With ten it is ten businesses' data.

If a second business cannot run on it without code changes, it is not a product yet.

Where do you start?

With an audit, before any new code. Read the code and the data, list what is specific to your business, and check for leaked keys and missing login checks. Then split the core from the custom parts, give each business its own data, make the connections swappable, and put one real second business on it.

• Audit: the code, the data, and a security check
• Split: what every business needs, and what only yours needs
• Separate: each business gets its own data and its own connectors
• Pilot: a second business runs on it, and every code change it needs is a gap to close

The pilot is the honest test. Whatever the second business asks you to change in the code is what still stands between your app and a product.

This is the work behind App to Product on our Services page. If another business has already asked for your app, send us what you built and who asked, at hello@99six.dev.

If that is your week too, thirty minutes is enough to find out.

Book 30 minutes →