There are three ways a booking reaches you, and they cost completely different things — one costs commission, one costs a fixed fee, and one costs the attention of whoever is standing nearest the phone at eight on a Friday. Most restaurants end up with all three by accident and have never compared them, which is how a small place ends up paying more per booked cover than its own dessert makes.
☎️ The three channels, honestly compared
| Channel | What it costs | Who owns the customer | What it is good at |
| The phone | Staff time, at the worst hour | You | Regulars, big groups, anything unusual |
| Your own site | A flat monthly fee, often small | You | People who already decided to come to you |
| A booking platform | Commission per seated cover | The platform | Being found by people who do not know you |
Read the third column before the second, because it is the one that decides the strategy. This is the same distinction the delivery course drew between a third-party channel and your own: a platform rents you its customers, and your own site serves the customers you already have. Both are worth having; paying platform prices for the second job is the mistake.
💰 What a platform actually costs
Do the arithmetic with your own numbers. In the example restaurant, roughly forty percent of the two thousand monthly covers arrive booked through a platform — eight hundred covers — at a typical fee of $1.50 per seated cover. That is $1,200 a month, or about 2.4% of sales. The same restaurant's own booking widget, on its own site, costs about $80 a month flat, whatever volume goes through it.
Now the uncomfortable question, and it is the one that decides whether that $1,200 is a bargain or a leak: how many of those eight hundred covers were people who already knew you? If seven hundred of them were regulars who typed your name into a search box and clicked the first booking button they saw, you paid roughly $1,050 that month for the privilege of taking bookings from your own customers. The hundred genuinely new ones cost you the rest — and those hundred may well be worth far more than $150, which is exactly why the answer is usually "keep it, but stop feeding it".
How to tell the difference is not sophisticated: ask the platform for the share of your bookings that are new-to-you diners, and compare the names against your own list, which the customer data course taught you to keep. Then do the one thing almost nobody does — make your own channel easier to find than theirs. Your booking link at the top of your profile, on your site, in your replies, on the printed check. Every regular you move from the platform to your own site is pure margin and changes nothing for the guest.
🔌 What to check before you sign
Whatever you choose, these are the questions that turn into problems later, in roughly the order they bite.
- Does it write to one book, in real time? Two booking sources that do not see each other will double-book you within a month. This is the single most important question on the list.
- Can you control availability by slot? Closing eight-thirty when the kitchen is already stacked, holding tables back for walk-ins, blocking a section for a private group. Without this, the platform paces your night instead of you.
- Do you get the guest's contact details? Some platforms give you the name and little else. That decides whether you can build a list at all, which is the customer data course's whole subject.
- Does it talk to your POS? Useful for linking spend to bookings, and a criterion rather than a project — how the integration is set up is the POS course's territory.
- What happens on a bad night? Cancellations, no-show marking, moving a booking from a phone at the door. If it takes six taps, nobody will do it at half past eight.
- How do you leave? Contract length, notice period, and whether your booking history and your guest list come with you. Assume you will want to switch one day.
And test it the way it will be used: on a phone, standing up, with one hand, by the person who is actually at the door on a Friday — not on a laptop in a quiet office by you.
📵 The phone is still a channel
Online booking does not remove the phone; it changes who calls. What is left is the difficult and valuable end — large groups, special occasions, allergies, "can we bring a cake", the regular who always wants the corner table. Those calls decide relationships, and they arrive at the worst possible moment.
Three habits cover almost all of it. Somebody owns the phone during service, by name, and it is not whoever happens to be closest. There is a script: greeting, name, party size, time, phone number, one line about anything unusual, and a repeat-back of the details — the missing phone number is the single most common cause of an unrecoverable no-show. And voicemail actually works: a message that says when you will call back, and a call back that happens. A voicemail nobody clears is worse than no voicemail, because it is a booking that thinks it exists.
One more thing worth saying out loud: whatever comes in by phone has to go into the same book as everything else, immediately. A reservation on a sticky note is a table you are going to lose or double-sell, and it is where almost every ruined Friday begins.