The booking system you should not build

I take bookings for a living now, which still catches me off guard when I say it out loud. The studio out the back gets hired by the hour: photographers who want the north light, a couple who record a podcast on Thursdays, a woman who shoots ceramics and needs the room dead empty. Every one is a booking — a time held, a deposit taken, a reminder sent, and now and then a person who simply does not turn up. So when an advisory client asks whether they should build their own booking system, I am not answering in the abstract. I am answering as someone who spent a recent Tuesday chasing a no-show for a slot I could have sold twice.

The answer is buy. Almost always buy. The rest of this is about the “almost.”

The honest answer, and the narrow case for building

Most bookings are the same shape. A person looks at what is free, picks a time, and pays or half-pays to hold it. That is a solved problem, solved by dozens of companies, several of them Australian, all of them cheaper than a fortnight of a developer’s time. If your booking is that shape you are not a special case, however much it feels like your business is unlike anyone else’s. Everyone feels like the exception. At the level of “someone picks a time,” almost nobody is.

Building is defensible when the booking is the product and the logic genuinely does not exist off the shelf. Not “we do it slightly differently” — genuinely does not exist. I have seen exactly one client where it was the right call: an equipment-hire outfit whose gear had interlocking dependencies, where hiring item A on the Friday made items B and C unavailable to a second customer on the Monday because of transit and cleaning windows. No scheduling product modelled that, and getting it wrong cost them real money in double-booked kit. That is a build. “We want the confirmation email to sound like us” is not a build. That is a text field in something you already pay for.

The tell is whether the difficulty lives in your rules or your taste. Taste you can buy your way around. Rules no product understands, you sometimes have to write — but that is the story everyone tells themselves, and it is usually the first that is true.

The build is the cheap part

Here is the thing nobody costs properly. The build is a one-off; the maintenance is a decade. Software is not a bookshelf you assemble once, it is a lawn. Card networks change their rules, an operating system update breaks the SMS integration, Victoria goes on and off daylight saving and something that worked in June throws a wobbly in October. None of that is your booking logic. All of it lands on whoever owns the system, and if you built it, that is you, forever, or until you pay someone to relearn what you already forgot.

I was that person. Before all this I ran infrastructure for a freight company out in Laverton, fourteen years in IT before a Canon and a knitting mill got hold of me. We had a dispatch-and-booking system a contractor named Wayne had written some time around 2008. Wayne left in 2011. The system did not. It ran the business — every job, every slot, every driver — written by one person no longer reachable who had documented precisely nothing. His name was still in the code, a comment on the line that calculated availability: a little “// Wayne, do not touch this” that I did in fact eventually have to touch. There was a scheduled task on a server under a desk that nobody could explain and everyone was frightened to disable. For four years a good chunk of my job was keeping a dead man’s software breathing. He was not actually dead. He was in Perth. It felt the same from where I sat.

That is what building buys you: the thing that outlives the person who understood it, and then becomes the most expensive object in the business precisely because it works and nobody dares open it.

What an off-the-shelf tool has already solved

When people price a booking system they picture a calendar and a form. That is the visible five per cent. What you are really buying is the boring, regulated ninety-five per cent you would not think of until it bit you.

Payments first, and the part of payments you genuinely do not want: PCI scope. If a customer’s card number ever touches your server, you inherit a compliance obligation that is nobody’s idea of a Tuesday. Use a hosted checkout — Stripe, Square, Zeller — and the card details go to them, not you, and you stay in the lightest tier of PCI, the one where you attest that you never handle the numbers. Building your own payment capture to save a percentage point is how a two-person studio ends up responsible for a data breach. Do not.

Then the rest of the iceberg. Two-way calendar sync, so an online booking blocks out the calendar you actually live in and does not double-sell while you sleep. Reminder SMS, trivial until you are running a messaging gateway and eating the failed deliveries. Timezone handling, which in this country is its own small hell: Victoria observes daylight saving and Queensland does not, so a Melbourne studio taking a Brisbane client over summer has a one-hour trap sitting right there, and a decent product has already fallen into it so you do not have to. A decade of other people’s bug reports, already fixed. That is the actual product.

For a service business here the shortlist is not exotic. At the time I am writing this:

Tool	Suits	Rough cost
Square Appointments	Solo operators; ties to Square card payments	Free plan for a single location
Calendly / Acuity	Advisory, consults, anything “pick a time with me”	Low monthly per user
Timely	Salons, studios, treatment rooms	Tiered monthly
Cliniko	Allied health, where notes and privacy matter	Per-practitioner monthly

Prices move, so treat that as a direction, not a quote. The floor is roughly a couple of coffees a day, against a developer at a hundred-and-something an hour with no end date.

What I actually do for the studio

I do not run any of the fuller systems, because the studio does not need one. I use the middle path, and recommend it to more clients than I expected: a plain booking page for the times, wired to a Stripe payment link that takes a deposit. That is it. No dispatch system under a desk. No Wayne.

The deposit is the whole trick, and it is the part small operators resist because it feels rude. It is not rude. It is the difference between a booking and a wish. For the first stretch of hiring the studio out I took nothing up front, on trust, like a mug, and I lost roughly one session in five to people who booked a free hold and then had a better idea on the day. On a room that only earns when it is occupied, that is not a rounding error, that is the rent.

I now take a fifty-dollar deposit to hold a slot, redeemable against the hire, forfeit if you pull out inside twenty-four hours. No-shows went from one-in-five to something I have to think hard to remember the last of. Nothing about the customers changed; the friction changed. A person who has put fifty dollars down at ten at night turns up, or messages you like an adult to move it. Fifty is small enough a real client does not blink and large enough a maybe either commits or reveals themselves before they cost you the morning.

A word on the money rules, because clients always ask and half of what they “know” is wrong. Your deposits and hire fees count toward turnover, and once your GST turnover reaches the seventy-five-thousand-dollar threshold you must register for GST and charge it; below that it is optional. Keep your records for five years — the ATO is entitled to ask, and “it was in an app that shut down” is not a defence. And on surcharging: for now you may not add a card surcharge above what accepting that card actually costs you, but that arrangement has a use-by date. In March the Reserve Bank confirmed that card surcharging on eftpos, Mastercard and Visa — American Express too — is removed from the first of October, after which you fold the card fee into your prices or wear it. Which is the whole case for buying rather than building: someone paid to follow that change has been following it for months. You have a business to run.

One thing, then: the deposit does more for your bookings than any software you could build, and it costs you a Stripe link and the nerve to ask for fifty dollars.