For two years I ran a transportation company called Swoop. Every month, $5,000 left the account for the software that made it run. Not for drivers. Not for vans. Not for fuel. For software. This past week I rebuilt what that software did for us, in three days, and the new one costs about five dollars a month.
This is a post about what that gap actually means, and it is not the post you think it is. It is not "custom software always wins." Sometimes renting is the right answer and I will show you where the line sits. But when a bill gets big enough, the line moves, and most business owners never stop to check where they are standing.
What $5,000 a month bought
Swoop was non-emergency medical transport, right here in Wayne County. We got people to dialysis, to appointments, to the places they could not otherwise reach. I launched it in 2023 and ran it until 2025. There is more of that story on my about page, including the ribbon cutting and the news coverage.
The platform underneath it was Autofleet. It did what a ride operation needs. Took the booking, priced the trip, put it in front of a driver, tracked the vehicle, moved the money. It worked, and I am not here to trash it. It is real software built by real engineers, and it did the job we hired it for.
It also cost $5,000 every single month, whether we ran four hundred rides that month or forty. Sixty thousand dollars a year. And when Swoop closed, all of it vanished at once. Two years of payments and I owned nothing. No code, no database, no system sitting on a shelf. The account switched off and there was nothing left on my side of the table.
Why Swoop closed, since it matters
Worth saying plainly, because it is the part people skip when they tell these stories.
Swoop did not close because of the software bill. It closed because of the mix of work. Most of our volume came through community action taxi passes and Medicaid. That work matters, and I am glad we did it. It is also priced by somebody else, paid slowly, and thin to the point of barely being there. The private paid rides were the only genuinely profitable part of the operation, and there were not enough of them to carry the rest. So I shut it down about a year and a half ago.
The software did not kill the company. But sixty thousand a year is real weight on a business already carrying a heavy load, and it bought us nothing we got to keep. That is the part that stuck with me.
Three days
Woosh Delivery is still running, and people around here still need rides. So this past week I sat down and built the thing again, this time as something we own. It is live at rides.wooshdelivery.com.
What it does:
- Quotes a real fare off real driving distance before anyone pays. $15 base, $3 a mile after the first. No surge pricing, ever
- Takes the card. Round trips get a hold, and the exact final fare is captured at drop-off, not a guess
- Gives every driver a portal on a private link. No app to install, no password to lose, no training session
- Dispatches itself. A new booking finds a driver and texts them the ride without me touching anything
- Runs the wait timer, extra stops, tips, and a cash turn-in ledger that balances at the end of the day
- Keeps rider accounts and saved addresses, and gives whoever is waiting a live status page
Three days from an empty folder to taking real bookings. Forty-nine commits. That is not me being a genius, and I want to be careful here, because this is exactly where these stories usually start lying to you.
The honest reason it only took three days
I did not rebuild Autofleet. I rebuilt the slice of Autofleet we actually used.
That distinction is the whole thing. A platform sold to hundreds of fleets has to work for hundreds of fleets. Every state's compliance rules. Every vehicle type. Every billing arrangement. Multi-depot routing for an operation with nine hundred vans. Reporting for a finance department that exists. All of that is real engineering, and all of it is why the product costs what it costs.
My version has to work for exactly one operation, in one town, with the rules we actually follow. I got to delete about ninety percent of the problem before writing a line of code. That is not clever. It is just the advantage of building for one instead of for everyone, and it is available to any business that has outgrown the generic version of its own software.
The other honest piece: I have been doing this daily for a while now, on the same stack, with tooling that has gotten genuinely good. The build is fast because the thousand small decisions were already made on the last ten projects.
What the five dollars actually is
The whole thing runs on Cloudflare. The code, the database, the scheduled jobs that sweep for unassigned rides, the hosting, all of it sits on their paid plan at five dollars a month. That is not a promotional rate or a trick. That is what this class of infrastructure costs now, and almost nobody outside the industry has noticed.
To be straight with you: card processing and text messaging are separate and priced by usage, the same as they would be on any platform. Those follow the volume. The number that went from $5,000 to $5 is the software itself, the part that used to be a flat monthly tax on the business whether it was busy or dead.
When you should absolutely keep renting
Here is the part most people selling custom software will not tell you.
If you are paying $50 a month for a tool that works, keep paying it. A custom build has to be worth building, and at $600 a year it almost never is. You would spend more of your own time managing the project than you would ever save. Rent it, use it, get on with your day.
The math changes on a curve, and it is worth knowing roughly where you are on it. A few hundred dollars a month is the gray zone, worth a conversation if the tool also fits your business badly. Past a thousand a month, it is worth a serious look. At five thousand a month you are not buying software anymore, you are leasing it at a price where owning becomes the obvious move, and every year you wait is another year of payments you keep nothing from.
The other trigger has nothing to do with price. It is when you are bending your business to fit the software. When your team keeps a spreadsheet on the side because the platform will not do the one thing you need. When you say "the system won't let us" to a customer. That is a real cost, it just does not show up on a bill.
The question worth asking
Pull up whatever you pay every month to run your business. Not the fun stuff. The operational software, the booking tool, the scheduler, the platform your whole day runs through.
Add it up for a year. Then ask what you would own if you stopped paying tomorrow.
For most business owners the answer is nothing, and they have never once done that arithmetic, because each individual bill was reasonable on the day it started. That was me for two years, and I did not run the numbers until the company was already closed and I had nothing to show for sixty thousand dollars.
Now I run a ride platform that does what the expensive one did for us, and it belongs to me. Nobody can raise the price on it. Nobody can sunset it, change the terms, or decide my business is no longer a good fit for their roadmap. That is worth more than the money.