The third bill.

I sell a competing platform. I have a horse in this race, and the horse has a name.

So read this letter the way you’d read a restaurant review written by the chef across the road: suspiciously, with your own contract open beside it. That isn’t a disclaimer. It’s the instruction. Nothing below names a vendor, and nothing needs to. Every mechanism in it is a clause you can find in your own paperwork by Thursday.

For six letters I’ve stayed inside the room: the seat, the chatbot, the leaks, the breakages, the tables everyone fights over, and last week the reviews. All of it front of house. This week I’m going round the back, to the office with the printer that only works for the GM, where a different story gets told once a month. It arrives as a PDF. It’s the booking platform’s invoice, and it’s the least-read document in hospitality, narrowly beating the allergen matrix.

It deserves more attention than that. There are three bills in it, and only one of them has a total.

The first bill: the subscription

This is the one everybody knows. A monthly figure, a line on the P&L, agreed once and renewed without ceremony. It’s visible, predictable, and the smallest of the three. Vendors will happily discuss it at length, which is usually a sign.

The second bill: the per-cover fee

Some contracts charge per guest seated through the platform. Sometimes that means only the guests the vendor’s marketplace sent you. Sometimes, if you read the small print, it also means guests who found you entirely on their own and simply pressed the button on your website. It’s worth knowing which yours is. Several operators I’ve asked didn’t.

Suppose the fee is small. Call it a pound-fifty, or seven-odd dirhams, though yours may differ. Now do the thing nobody does and multiply it out. A venue seating 800 booked covers a week, which isn’t a large venue, pays £1,200 a week. That’s £62,400 a year, before the subscription, because a guest pressed a button.

Then look at when it bills. The per-cover fee peaks on your fully booked Friday, the night the room would have filled anyway on reputation, the weather and the fact that it’s Friday. It idles on your dead Tuesday, the one night you’d welcome some help. (The manager’s off on Tuesdays. So, it turns out, is the fee.)

It’s a tax on your own success, billed as a service.

Enter Procurement

Which brings me to the newest member of the cast.

Procurement has never seen the terrace. Procurement has seen a spreadsheet of the terrace, and was very pleased with it. Procurement met the vendor at a conference (somebody said "game changer"; somebody always does) and negotiated the group deal at head office over two excellent lunches. Procurement would like you to know it’s a very good rate.

Group rate, that.

The deal was negotiated at head office. It gets paid for, per cover, on the terrace.

Nobody in the building where the fee is incurred was in the room where it was signed. That isn’t a villain story. Procurement is doing his job, and by the measure he was given he’s doing it rather well. It’s a measurement story. The deal was judged on the subscription because the subscription is the number on the page. The other two bills were never on the page.

The third bill: paid in guests

This one doesn’t arrive in an envelope.

Somewhere in your contract is a clause about data, and it decides who owns the diner. Not the booking: the diner. Her name, her number, her four visits, her allergy, the anniversary she mentioned, the terrace she always asks for. She’s the guest from 7:42 on a Friday, the one this newsletter has spent six weeks trying to get to your table.

Read the clause and ask three things. Can you export all of it, whenever you like, in a format another system can read? Can the vendor use your guests to market other venues, occasionally including the one across the road? And when you leave, what happens to her?

In most venues, who owns the guest was never decided. It was signed.

Earlier this month I went looking for why venues leave the booking platforms they’re on. I didn’t ask the vendors; I read what operators say about them in public, when they’re cross enough to write it down. I’d expected features. What came up again and again was the commission on every booking, a booking page that never looked like their restaurant, and nobody picking up when it broke. None of the three was about what the software could do.

The venues that do decide to leave then meet the last mechanism: the manufactured exit. It’s the quietest of the lot. Nobody tells you that you can’t leave. The exit just costs something: an export fee, a notice period that happens to close the week after auto-renewal, a "migration service" billed by the hour, your data handed back in a format nothing else reads. None of it is a wall. It’s a slope, and on a slope most people stop walking.

The bit where I have to be fair

This letter owes an honest part, and I said I’d write them that way.

Switching is never free, including to me. Moving platforms costs a fortnight of a manager’s attention, retraining every host, and a service or two where everyone is slightly slower and pretending they aren’t. Anyone who tells you it’s painless is selling the switch, not the platform. Some of you will read this, do the sum, and decide your current deal is fine. Some of you will be right.

But one test cuts through every sales deck, mine included:

A vendor confident in their product makes leaving easy, and stays anyway.

So ask your vendor the three questions. Then, and I mean this, ask me the same three. Horse, name, race. You were warned in the first line.

Next week: the AI-ready self-assessment. The questions I ask when I audit a venue, handed over so you can answer them about your own in an afternoon, without buying anything from anyone. Me included.

— B

Disagree with any of this? Message me. I answer.

Bhrij Patel is the founder of RAYN, an AI-native hospitality intelligence platform, built after years operating restaurants from quick service to fine dining. For an honest, vendor-independent read on where AI fits your operation, ask about the AI Readiness Audit