Systems

When one client’s system turns into a product.

We built an enquiry app for a cleaning company. It worked well enough that we rebuilt it as something we sell. That is not an accident, and it is not luck either.

18 November 2026

Most bespoke software is built for one business and dies with that business. It gets used for three years, the person who commissioned it moves on, and eventually somebody replaces it with an off-the-shelf thing that does eighty per cent of the job.

Occasionally one goes the other way.

What we actually built

A cleaning business needed enquiries to stop falling down the back of the sofa. Someone would message at eleven at night asking whether next Thursday was free, and the answer depended on a diary in somebody’s head.

So we built the thing that answered that question. Availability, an enquiry flow that felt like the apps people already use rather than a form from 2009, and a route straight into WhatsApp, because that is where the conversation was going to happen anyway.

It was built for one company, with their name on it, for their jobs. That is the only way we know how to build something people actually use.

Then it kept working

The tell was not that it launched. Everything launches. The tell was that six months later it was still being used, by people who had no obligation to use it, because going round it was more effort than using it.

That is a rarer signal than it sounds. Most internal tools get quietly abandoned in favour of a WhatsApp group and a spreadsheet, and nobody tells the people who built it.

So we rebuilt it as a product. It is now G Flow — booking software sold to cleaning businesses, built on top of what worked for one of them.

What that says about building bespoke

The lesson is not that every client system should become a product. Most should not, and trying to make them is how you end up with software that is vague enough to suit everybody and useful to nobody.

The lesson is the other way round. The reason it was worth turning into a product is that it was built for one business with one specific problem, by people who went and watched how that business actually worked. The specificity is what made it good. Generalising came afterwards, and only because the specific version had earned it.

If we had set out to build booking software for cleaning companies in general, we would have built something worse.

What it means if you are the client

Two honest things.

The first is that a system built around your business will be built around your business. Yours is the version that is exactly right. Anything we learn from it that becomes a product is our problem to solve, not something you pay for.

The second is that we are building these things with our own money on the line elsewhere, which changes how we build. We are not handing over a prototype and walking away, because we have found out the hard way what it costs when software has to keep working on a Tuesday morning with nobody around to restart it.

The uncomfortable version

Plenty of agencies will build you a bespoke system. Fewer have had to live with one.

The difference shows up in the boring parts — what happens when the internet drops, when two people book the same slot, when somebody types their phone number with spaces in it. None of that is in the pitch. All of it decides whether the thing is still being used in a year.

If this is your problem

All insights

Rather just talk it through?

Tell us where the business is now and we will tell you what we would actually do.

Start a conversation