Every restaurant that has ever tried to leave a booking portal has run the same internal argument. The commission is painful, but the portal delivers covers, and nobody wants to be the person who turned off a channel and watched Friday go quiet.
That fear is reasonable and it is also the entire business model of the aggregator. This article is about reducing commission dependency deliberately rather than dramatically, which is the only version that works.
First, understand what you are actually buying
You are buying two different things from a portal and they have very different values.
Discovery. A guest who did not know your restaurant existed found it through the platform. This is genuinely valuable and it is what commission is meant to pay for.
Convenience. A guest who already knew your restaurant, already intended to eat there, and used the portal because it was the easiest booking route in front of them. You paid commission to serve a customer you had already won.
The second category is usually larger than operators expect. Every regular who books through the portal out of habit is a commission charge on a relationship you built and maintain yourself.
Your first job is not to leave the platform. It is to move the second category off it.
Step one: measure the split before changing anything
You cannot manage this without numbers, and most operations do not have them.
Pull three months of bookings and split them by source. Then, for portal bookings, check how many of those guest phone numbers appear more than once. A guest booking through a portal for the third time is not discovery, it is habit, and it is the population you can move without any risk to cover volume.
Also calculate your actual blended commission rate as a percentage of covered revenue rather than as a per cover fee. The per cover number sounds small. The annual figure against revenue is the one that gets attention in a management meeting.
This measurement is also the baseline you will report against later, so do it properly before you change anything.
Step two: build a direct booking route that is genuinely better
The reason guests use portals is that the alternative is often worse. If your direct route is a contact form that generates an email somebody checks in the morning, guests will keep using the portal and they will be right to.
A direct booking route that competes needs four things:
Real time availability. Not a request form. If the guest cannot see that eight thirty on Friday is open and confirm it immediately, you have not built a booking system, you have built a queue.
No account creation. Name, mobile number, party size, done. Every additional field costs you completions.
Immediate confirmation on the channel the guest uses. In this market that means WhatsApp, not email. A confirmation that arrives where the guest already is feels like a real reservation. One that lands in a promotions tab does not.
Mobile first. The overwhelming majority of restaurant bookings in the UAE happen on a phone, frequently while the guest is doing something else.
Get those four right and your direct route is better than the portal experience, not merely cheaper for you. That distinction matters, because guests do not care about your commission rate.
Step three: put the direct route where the intent already is
Discovery happens in specific places. Booking should happen in the same place, not two clicks later.
Instagram bio and story link. For restaurants in Dubai, Instagram is frequently the discovery channel. If the bio link goes to a homepage rather than a booking page, you are losing bookings at the exact moment of intent.
WhatsApp. A click to WhatsApp link that opens a conversation and leads to a booking is an extremely low friction route in this market, and it doubles as a consented contact for later.
Google Business Profile. Make sure the reservation link points at your own booking page rather than a portal, and keep your hours accurate. This is free and frequently misconfigured.
A QR code on the table. The guests sitting in your restaurant right now are your highest converting audience for a return visit. A table QR that books the next visit or enrols them in your loyalty programme captures them at peak satisfaction.
Your own website. Booking above the fold, on every page, not buried under a menu tab.
Every one of those routes should write into the same guest record. A booking that does not build your database is a missed opportunity even when it saves you commission.
Step four: capture the guest, not just the booking
This is the part that turns a cost saving into a compounding asset.
Each direct booking should create or update a guest profile holding visit history, spend where you can capture it, outlet preference, occasion dates, dietary notes and explicit communication consent. That profile is what lets you bring the guest back without paying anyone.
Consent is not optional. Capture it explicitly at the point of booking, record what the guest agreed to and when, and honour withdrawal. This is a requirement under UAE data protection law and a practical necessity if you intend to message guests on WhatsApp, where Meta’s platform rules require opt in as well.
The distinction that matters: on a portal, the platform owns that relationship and rents it back to you. In your own system, it is yours, and it is exportable. We explain how the data ownership question should be handled contractually in our comparison of platform options for UAE hotel food and beverage.
Step five: give guests a reason to book direct next time
Better booking experience gets you parity. A reason gets you the switch.
Loyalty is the most durable version. Points that only accrue on direct bookings, redeemable across your outlets, create a standing reason to skip the portal. It is not a discount, it is a relationship, and it does not train guests to wait for offers the way discounting does.
Other mechanics that work:
- Priority access to peak slots for members
- Occasion recognition, since a birthday remembered is worth more than a percentage off
- First access to new menus or events
- A small welcome benefit at enrolment to make the first sign up worthwhile
If you run several outlets, the loyalty currency should work across all of them. That multiplies the reason to enrol and it is the mechanic portals cannot replicate. The operational detail is covered in running one loyalty programme across multiple restaurants.
Step six: reduce dependency in a controlled way
Do not switch the portal off. Reduce your reliance on it and let the numbers tell you when to change the arrangement.
A sensible sequence over roughly six months:
Month one. Measure. Establish the source split, the repeat portal booker population and your blended commission rate.
Months two and three. Launch the direct route and place it everywhere intent occurs. Change nothing about your portal allocation. You are building an alternative, not withdrawing a channel.
Months three and four. Launch loyalty. Enrol aggressively at the table, where your guests are already happy and already present.
Months four to six. Begin messaging your consented database directly. Confirmations, reminders that reduce no shows, occasion messages, and offers for quiet periods.
Month six. Review. Look at what share of covers now arrive direct, and specifically at how many previously portal booking guests now book with you. Then decide whether to reduce portal inventory, renegotiate the rate using your new direct volume as leverage, or leave it as it is because the discovery it delivers still pays.
That last option is a legitimate outcome. The goal is that the portal becomes a discovery channel you choose to use rather than a dependency you cannot exit.
What to expect, honestly
Direct booking share moves gradually. A well executed programme typically shifts a meaningful minority of bookings in the first six months, and the shift continues as your database grows. Anyone promising you a majority in a quarter is selling something.
The compounding effect is where the value sits. Each guest captured directly is contactable forever at no marginal cost beyond the message. The portal booking is a transaction. The direct booking is an asset.
The mistakes that cost people the most
Turning the portal off first. Build the alternative, then reduce. Every operator who reverses this order ends up back on the platform at a worse rate.
A direct route that is worse than the portal. If your booking flow is slower or less certain, guests will not use it however much you would prefer they did.
Not capturing consent. A database you cannot legally message is a list, not an asset.
Discounting instead of relationship building. Price promotions to drive direct bookings train guests to wait for discounts. Loyalty and recognition do not.
Letting bookings sit outside the guest record. If the reservation does not build the profile, you have saved commission and gained nothing durable.
Where to start this week
Pull three months of bookings and split them by source. Count how many portal bookings came from repeat phone numbers. Multiply your commission rate by that subset.
That number is what you are currently paying to serve guests you had already won. It is usually enough to get the project approved.
If you want help building the direct route and the guest database behind it, we build this for restaurants and hotel food and beverage operations in the UAE, and we are based in Dubai.
