Aggregator Economics 2026: What Delivery Really Costs in the UAE, KSA and India

Delivery added revenue for almost every restaurant in the UAE — and quietly compressed margin for many of them. An owner looks at a busy delivery screen and assumes the orders are helping. Sometimes they are. Sometimes each one hands back more than it brings in, and the only way to know is to read the delivery channel’s margin by itself.

This is the operator’s view of how aggregator economics actually work in 2026 — what comes off the top of an order, how the tiers are structured, how a per-order rate turns into a share of your whole business, and the weekly read that keeps the channel honest. The figures here are teaching bands and illustrative arithmetic, not quotes: your own contract and your own weekly statement are the numbers that count.

How does aggregator commission actually work?

When a customer orders through Talabat, Deliveroo or any other aggregator, the platform takes a commission — a share of the order value — in exchange for the marketplace, the demand, the payment handling and usually the delivery itself. As a teaching band, commission typically runs between 15% and 30% of order value, and where a given restaurant sits inside that band moves with category, contract, delivery mode and the extras signed up for: paid placement, exclusivity arrangements, whether the platform’s riders or your own carry the food. When the Khaleej Times surveyed Dubai operators in 2020, the quoted range was 25–30%, “up to 35%” with everything stacked on top. Confirm your own contract terms — a band is not your rate, and the platforms’ published partner terms deliberately leave the percentage to the individual agreement: Deliveroo’s UAE partner terms, for instance, define the partner payment as the menu-items amount less “the Fees applicable in the Agreement”, calculated weekly.

Commission comes off the top of the order, before you have paid for a single ingredient. So the question is never “what is the commission?” — it is “what is left, per order, after commission and everything else delivery costs me?” Best answered by walking one order all the way down.

The full order P&L: one delivery order, walked to the end

Take a single order and follow it to the last dirham. Illustrative, round numbers — your own statement gives the real ones:

Line	Basis	AED	% of ticket
Menu price on the app	—	60.00	100.0%
Commission	25% — upper half of the 15–30% band, common for full-service tiers where the platform’s riders deliver	−15.00	25.0%
Packaging	Box, bag, sleeve, cutlery, seal	−3.50	5.8%
Promotion and ads share	The discount that won the order plus the day’s sponsored placement spread across its orders	−6.00	10.0%
Food cost	30% of menu price, inside the 28–32% band	−18.00	30.0%
Contribution before labour, rent, utilities		17.50	29.2%
Labour	Middle of the 25–30% band, call it 27.5%	−16.50	27.5%
Left before rent, utilities, repairs		1.00	1.7%

Check it: 60 − 15 − 3.50 − 6 − 18 = 17.50, which is 29.2% of the ticket. Set labour against it at 27.5% — AED 16.50 — and the order is holding AED 1.00. One dirham, before rent, electricity or a single repair — and only if nothing was remade, refunded or returned. That is the arithmetic hiding inside a “busy” delivery night, and why the channel must be read on its own: a healthy dine-in room can subsidise this quietly for months.

Why does a dine-in-profitable dish lose money on delivery?

A dish is priced for the table, where the only deductions are food cost and the cover’s share of fixed costs. On delivery the same dish, at the same price, has to absorb a stack the dine-in price was never engineered for: a double-digit commission off the top, packaging on every order, and the promotion that won the click. On a discounted order with premium packaging and a top-of-band tier, contribution goes negative — you are paying for the privilege of fulfilling it. The menu price never changed; the channel changed everything around it. Packaging deserves its own P&L line for exactly this reason; the packaging economics read walks how to cost it per order the way you cost a recipe.

How are commission tiers typically structured — and what does a marketplace-only tier look like?

Commission is a menu of tiers, not one number, and in our experience across GCC operations the structure is consistent even where the rates differ:

Lever	What moves the rate	Direction
Delivery mode	Platform riders (full service) vs your own riders (marketplace only)	Full service sits at the top of the band; own-rider tiers materially lower
Exclusivity	Single-platform commitment vs multi-homing	Exclusivity typically buys a lower rate; multi-homing costs more per order but keeps demand options open
Marketing	Sponsored placement, promotion days, co-funded discounts	Stacks on top of commission — the effective take is commission plus all of it
Introductory terms	New-partner rates	Real but temporary; diarise the step-up date
Volume	Order history and a credible own-channel alternative	Tiers are renegotiated, not posted; operators with neither take the rack rate

One published example shows what a differently shaped fee stack does to the arithmetic. In October 2022, talabat and the Dubai Restaurants Group (now the UAE Restaurants Group) announced a Digital Growth Program for qualifying members — capped at the first 500, valid for two years — with commission of 5.3%, a delivery fee of AED 8.40 per order, and a 2% card or cash-handling fee. A dated, membership-conditional programme, not a rate anyone can assume today; but its structure — a low percentage plus a fixed per-order fee — is the teaching case for why ticket size decides everything:

Ticket	5.3% + AED 8.40 + 2%	As % of ticket	Flat 25% commission
AED 60 order	3.18 + 8.40 + 1.20 = 12.78	21.3%	15.00 (25.0%)
AED 30 order	1.59 + 8.40 + 0.60 = 10.59	35.3%	7.50 (25.0%)

On the AED 60 order the fixed-fee structure is cheaper than a flat 25% (12.78 against 15.00). On an AED 30 order it is far more expensive (10.59 against 7.50) — the AED 8.40 delivery fee alone is 28% of a small ticket. The lesson generalises to every fee stack with a per-order component: a percentage scales with the ticket; a fixed fee punishes the small one. Only your own average ticket tells you which suits you. The discipline is to know, in writing, which tier you are on, what stacks on top of it, and when it changes.

How does a per-order rate become a share of my whole business?

Two percentages wear the same name. The per-order commission is the rate on the app. The commission as a share of total revenue — the line on the Restaurant Operating Index, banded at 3–6% — is that rate multiplied by how much of your business runs through the platforms:

Delivery share of revenue	at 15% commission	at 25%	at 30%
10%	1.5%	2.5%	3.0%
20%	3.0%	5.0%	6.0%
30%	4.5%	7.5%	9.0%
40%	6.0%	10.0%	12.0%

A restaurant doing a fifth of its business through the apps at 25% sits at 5% of revenue — inside the band. Let delivery grow to 40% of revenue at the same rate and commission alone takes 10% of everything you sell, more than most operators pay in rent. Nothing on the menu changed; the mix did. This is why our rescue check reads a per-order commission above 25% as structural rather than drift, and why the Profit Leak Audit asks for delivery commission as a share of revenue beside food, labour and rent: it is one of the four lines that decide margin, and the one that grows without anyone deciding it should.

Read the margin by channel — and cost the two quiet leaks

Stop looking at one blended margin and start reading margin by channel. Most P&Ls mix dine-in and delivery into one revenue and one food-cost line, which is exactly what hides a delivery problem — strong dine-in margin papers over weak delivery margin until the blended number quietly sags. Per delivery order, the honest calculation is:

Order value − commission − packaging − promotion/ads − food cost = true delivery contribution

Run that on your actual orders, not hopeful ones, and watch the two per-order costs that escape scrutiny: packaging, which scales with every order rather than with revenue — spec it to the dish, cost it per order, resist the most premium box on the shelf; and promotion drag, which stacks on top of commission and packaging on the same ticket, so a modest-looking discount can be the difference between positive and negative contribution. Treat every offer as a margin decision with the post-commission arithmetic in front of you — never a reflex.

The same channel-margin discipline sits at the heart of the cloud-kitchen model, where almost every order is delivery and there is no dine-in margin to hide behind. The Delivery Margin Recovery tool separates your delivery contribution from dine-in, commission included.

Own channel versus marketplace: decide the mix

Aggregators bring reach you cannot buy elsewhere; the mistake is letting them own all of it. Every order through your own website, app or WhatsApp pays payment processing and your delivery arrangement instead of full commission, and the blend moves fast: if three orders in ten shift to an own channel at negligible commission, the blended commission on those ten falls from 25% to 17.5% (7 × 25% ÷ 10) — a 7.5-point structural improvement without renegotiating anything.

Put the same AED 60 order through an own channel and the stack changes shape. Illustratively — payment processing at 2%, your own rider or a courier at a flat AED 9, the same packaging and food cost, no platform promotion:

Line	Marketplace (25%)	Own channel
Menu price	60.00	60.00
Commission / payment processing	−15.00	−1.20
Delivery cost	(included)	−9.00
Packaging	−3.50	−3.50
Promotion and ads	−6.00	0.00
Food cost (30%)	−18.00	−18.00
Contribution before labour	17.50 (29.2%)	28.30 (47.2%)

Check it: 60 − 1.20 − 9 − 3.50 − 18 = 28.30. The own-channel order keeps AED 10.80 more than the marketplace order. That is not an argument for zero aggregator volume — the marketplace order exists because the marketplace found the customer. It is an argument for moving the repeat customer, who no longer needs finding, onto the channel that keeps 47% instead of 29%.

The practical playbook is unglamorous: reorder cards and QR inserts in every bag, consented repeat customers captured onto direct ordering, own-channel pricing or bundles kept slightly favourable, and the marketplaces left to do what they are genuinely good at — discovery and incremental demand. The target is a deliberate mix, chosen with the channel margins in front of you, instead of a default the apps chose for you.

UAE, KSA and India: the same machine on different settings

The order P&L above works in every market; what shifts is the setting on each dial. Framed as typical structures, from our experience across GCC and India operations — not statistics:

• UAE: a concentrated marketplace — Talabat and Deliveroo carry most third-party demand — with high delivery penetration and tickets that absorb the cost stack better than most markets. The trap is complacency at the top of the band because the room is busy.
• KSA: platforms such as Jahez and HungerStation shape the market, and geography does the rest — Riyadh and Jeddah distances make delivery zones and own-rider tiers a bigger part of the negotiation, and self-delivery is a live option for more operators than in the UAE.
• India: Swiggy and Zomato dominate and tickets are lower, so fixed per-order costs like packaging weigh more — the small-ticket effect in the fee-stack table above — discount culture runs deeper, and visibility is increasingly ads-led, so the effective take that matters is commission plus ads plus discount funding, which operators typically describe as running well above the headline rate.

The lesson is the same in all three: model the channel on your market’s real settings, never import another market’s assumptions, and never trust a headline rate anywhere.

When is volume bought at negative margin — legitimately?

Sometimes an operator runs delivery orders at a loss on purpose — and sometimes it only looks like purpose. In our experience the distinction separates a strategy from a leak:

• Legitimate, briefly: a launch window where discounted trial buys first orders and reviews — with a budget, a target and an end date.
• Legitimate, narrowly: marginal capacity. When the kitchen and team are already paid for and idle, an order whose contribution covers its variable costs but not its full-cost share can be rational — if you know the number and chose it. Break-even tells you where that full-cost share sits.
• Illegitimate, always: drift. Promo-stacked orders with negative contribution — below variable cost — running indefinitely because nobody reads the channel P&L. That is not volume; that is paying the platform for the privilege of being busy.

The rule: negative-margin volume is a marketing spend with a budget and an end date, or it is a leak. There is no third category.

The weekly aggregator read

Everything above compresses into one habit. Once a week, same day, the manager pulls the platform statements in a fixed thirty minutes and reconciles the channel. Illustratively, one platform, one week:

Line	AED
Gross order value	40,000
Commission taken (contracted 25%)	−10,000
Sponsored placement and ads	−1,600
Co-funded promotions	−1,200
Refunds, adjustments, chargebacks	−400
Net payout	26,800
Effective take = (40,000 − 26,800) ÷ 40,000	33.0%

The contract said 25%. The statement says 33%. That eight-point gap — ads, promotions and refunds, the stack on top — is the number that governs the channel, and it only exists if someone computes it. Check the commission line against the contracted tier every week (a step-up from expired introductory terms shows up here first), then watch the effective take and the promo share as trends, not events: an ads creep or a refund pattern appears here weeks before it reaches the monthly P&L.

The Aggregator Commission Tracker does this arithmetic for you — enter the week’s figures and it returns your effective take and where the stack is heaviest, free and confidential. Five nights of the aggregator cut is the same arithmetic told from the pass. And if the wider numbers have already thinned — revenue holding, profit sagging — the delivery channel is one of the first places a turnaround looks, because it is where healthy-looking revenue most often hides an unhealthy margin.