Direct Restaurant Bookings vs Booking Marketplaces
Marketplaces genuinely sell covers you would not otherwise have sold, and they genuinely charge you for covers you would. Here is how to weigh the two honestly, and how to tell which kind of booking you are actually paying for.
The argument about marketplaces is usually conducted by people with something to sell. Marketplaces tell you they bring guests you would never have reached. Booking software vendors tell you the fees are a tax on customers you already had. Both are describing something real, and neither is describing your restaurant.
The useful question is not which side is right in general. It is what proportion of the covers you are paying commission on are genuinely new, and whether that proportion justifies what you hand over. That is answerable with your own data, and this article is about getting to the answer.
What a marketplace genuinely gives you
Start with the honest case for them, because operators who dismiss it tend to leave money on the table.
- Discovery you cannot replicate. Someone searching for late Thai food near a station at nine on a Tuesday is not searching for you. They are searching for a category, and a marketplace is where that search happens.
- Demand on your worst shifts. A quiet Monday with an empty room costs you the same in staff and rent as a full one. Covers that arrive at a fee are better than covers that do not arrive.
- Search visibility you would spend years earning. Marketplace listings rank for the category and neighbourhood terms an independent restaurant site rarely competes on.
- A working booking flow from day one. For a restaurant with no site worth booking through, a listing is functioning infrastructure that costs nothing up front.
- Trust for guests who do not know you. A first-time diner who has never heard of you may be more comfortable booking through a name they recognise.
The marketplace case is strongest when you are new
In your first year, when nobody is searching your name, discovery is worth almost any reasonable fee. The economics change as your own brand searches grow — which is exactly when most operators stop re-examining the arrangement.
What it costs you
The per-cover fee is the visible cost and usually the smallest one. Four others matter more over a couple of years.
- Fees on covers you already had. A regular who books through the app because it is on their phone costs you exactly the same as a first-time diner the platform found for you. The platform cannot tell the difference and does not need to.
- A guest record you do not own. You may see a name and a covers count. You often cannot email that guest, cannot see their full history with you, and cannot take them with you if you leave.
- Pressure towards discounting. Visibility on a marketplace is competitive, and the levers offered are usually price levers. Discount-led covers also tend to show the weakest completion rates and the thinnest spend.
- A listing that competes for your own name. Marketplace pages are heavily optimised. Searching your restaurant by name can surface the listing above your own site, alongside three competitors on the same page.
- Control of the messages. The confirmation and the reminder are the two communications a booked guest reliably opens, and on a marketplace booking they carry the platform’s branding, not yours.
What direct bookings give you
A direct booking is one made through your own site, on your own domain, into your own system. The margin is the obvious benefit and again the least interesting one.
You get the full guest record: contact details, history, allergies, the fact that they came for an anniversary last March. You control the confirmation and the reminder, which is where no-show rates are actually won. You can change your rules — turn times, party-size ceilings, last seating — without asking permission. And nothing you spend building your own brand gets intercepted at the last step by a page listing your competitors.
The honest downside is that direct bookings require demand you have already created. A booking widget on a site nobody visits produces nothing. Direct is a conversion channel; it is not, by itself, a discovery channel.
Side by side
New demand, or interception
This is the question that decides everything, and it is answerable without any special tooling. You are trying to separate two kinds of marketplace booking: the guest who found you there, and the guest who already intended to eat with you and used the app because it was convenient.
Four checks get you close enough to act on.
- Look at the new-versus-returning split in the platform’s own reporting. A listing where a large share of covers are returning guests is mostly billing you for your own regulars.
- Compare the days. Genuine discovery skews towards your quiet shifts, because that is when you have availability to be discovered on. A marketplace that mainly fills your Saturday 8pm — the slot that sells itself — is intercepting, not generating.
- Watch what happens when you cut visibility. Reduce your promoted inventory or pull back the discounted slots for a month. If your total covers hold roughly steady while marketplace covers fall, those bookings were coming to you anyway.
- Check your own brand searches. If people are searching your restaurant by name in growing numbers, that demand is yours. Anything that captures it downstream is charging you for delivery, not for the customer.
A month is not long enough to judge the cut-back test
Guests who cannot find you on a platform do not stop wanting dinner immediately; they drift back through search over weeks. Run the test for at least two months, and do not run it across a holiday period or a menu change.
Run both, but make direct the default
For most independents the answer is not to leave the marketplace. It is to stop treating it as your booking system and start treating it as a paid acquisition channel with a cost per booking, judged the same way you would judge an advert.
- Make your own site the fastest way to book: a booking button in the header of every page, and the booking flow on the homepage, the menu page and the contact page.
- Point every channel you control at your own site — Google Business Profile, Instagram and Facebook bios, your email signature, printed QR codes, the footer of every newsletter.
- Keep one book. Whichever system holds real availability must hold every channel, so a marketplace cover and a phone cover cannot claim the same table.
- Give the marketplace your genuinely hard-to-sell inventory: early sittings, quiet weeknights, the shoulder hours. Hold your prime slots for direct.
- Put your restaurant name, in full, on every direct channel you own, so a guest who has decided on you can find you before they find a list.
- Review the arrangement every quarter against a number, not a feeling.
Never make direct booking the slower option
If your own booking flow is three taps longer than the app, guests will use the app and you will pay for the privilege. Direct only wins on convenience, so it has to actually be more convenient.
Measuring the split
Set one number in front of yourself every month: the share of covers that arrived direct. Then set a second: what a marketplace cover costs you in fees, divided by the covers you believe were genuinely new. That second figure is your real cost per acquired guest, and it is usually several times the headline fee.
On the direct side, tracking links let you attribute bookings back to the channel that produced them, so "direct" stops being one undifferentiated lump. In Seatingly, bookings that arrive without a tracking link are labelled Direct/Unknown — a large share there is normal, since it includes walk-ins, phone bookings and regulars going straight to your site. Attribution runs on a 30-day window, so judge a campaign a month after it ends rather than the morning after. Create marketing tracking links covers the setup, and Measure booking ROI in Marketing Insights covers reading cost per booking once the data is in.
A reasonable ambition for an established independent is a clear majority of covers arriving direct within a year, with the marketplace kept deliberately for the shifts you struggle to fill. Whether you get there matters less than knowing the split at all. Most operators discover, on measuring it for the first time, that they were paying a discovery fee on guests who had known their name for years.
If this is prompting a wider rethink of what you are running, How to Choose a Restaurant Reservation System works through what to insist on and what to ignore.
