Fantastic Services didn’t build ServiceOS because they wanted software. They built it because scale broke everything else.
They started with cleaning, two laptops, and a shared mobile phone. Bookings ran through Google Docs. Then the browser started crashing — too many bookings in one day.
First the schedule held a week. Then a day. Then one day broke the browser.
That’s what scale does. It rarely breaks dramatically. It breaks because the informal system that got you here can’t carry what comes next.
The owner’s memory. WhatsApp threads. Spreadsheets. The “just ask Dave, he knows that site” workaround.
At a certain size, you need an operating layer.
That’s what ServiceOS became: scheduling, pricing, customer management, billing, reporting, and operational control. A system to hold the business together as complexity grew.
Fantastic Services built ServiceOS because scale required it.
I’m looking at the same problem from a different angle. Before I use AI in my first cleaning acquisition, I need to answer a more basic question:
Is there clean, usable operational data behind this business? Or is it still held together by memory, WhatsApp, and the people who’ve always known how it works?
The real lesson isn’t software. It’s what breaks without a system.
The easy takeaway from Fantastic Services is: “they built software”.
The real lesson is that repeated service problems need a system.
Calls need answering. Quotes need pricing. Jobs need scheduling. Teams need routing. Quality needs proving. Management needs visibility.
When those things sit in people’s heads, the business can still work. That is how most founder-led service businesses get built. Win work, solve problems, keep clients happy, and rely on a small number of good people to remember what matters.
As you grow, that same strength becomes the constraint.
The founder knows too much. The admin person becomes the system. The business scales revenue, but not control.
That’s where a big acquisition risk sits.
Not in whether the business has clients.
Whether the operating model can survive without the founder holding it together.
The accounts show what happened. The operating data shows what you’re really buying.
A cleaning business isn’t just a client list.
It’s a network of sites, routes, labour blocks, time windows, access rules, keyholding arrangements, cleaning specifications, informal quality checks, pricing assumptions, complaint history, travel patterns, and scope creep.
That network is either visible or invisible.
If it is visible, you can improve it.
If it is invisible, you inherit fragility.
That’s one of the biggest shifts in how I think about buying cleaning businesses now.
The accounts tell you what happened. The operating network tells you whether the business can be improved.
Two businesses. Same revenue. One has an operating layer. One has the seller’s memory. Different asset entirely.
Same revenue. Different risk. Different transferability. Different value.
This echoes a conversation I had with Avin Rabheru, founder and CEO of Housekeep. His point was simple: operational control in cleaning shows up in routing, density, job data, digital communication, and proof of service.
Operational control is a data problem, not a tech one.