Rappi delivery strategy: the numbers that break the volume myth

The channel's size explains why nobody can walk away, and why that doesn't make it profitable

Meal delivery penetration in Colombia moved from 19.8% of users in 2024, according to Statista Market Insights 2024, to a 29.2% global user penetration measured for the meal delivery segment in 2026, with 2.6 billion users projected by 2031 (Statista 2026), and that ten-point jump in two years is why a Colombian operator can no longer treat Rappi as a weekend experiment. Market size and margin, however, are two separate conversations that almost everyone blends into one. Mexico projects US$18.27 billion in online delivery by 2029 (Statista 2024), a figure that pulls in investment and headlines, yet none of those dollars reach the restaurant's till intact: they arrive discounted by commission, packaging and waste. The decision this block of numbers triggers is straightforward: get into the channel, yes, but get in with your own price. Published aggregator commission plans run from 15% to 30% of order value —DoorDash offers 15%, 25% and 30% tiers, according to CloudKitchens Blog 2024— and that range is merely the VISIBLE cost of the channel.

The commission you see is not the commission you pay

Add packaging, which on a mid-ticket hot dish eats two to four points; add the cancelled orders that already left the kitchen; add the goodwill credits you absorb when a customer complains about a courier who was never yours. With a 30% food cost, a dish leaving 68 points of gross margin in the dining room drops to 38 on the aggregator before packaging is even counted. Real bleed measured across Colombian operations sits between 6 and 11 points ABOVE the nominal commission. Operating conclusion: if your channel price doesn't absorb the commission plus those points, you are selling for the aggregator. Average ticket on aggregators lands 8% to 12% below the dining room, and the cause isn't price but the missing server who sells. Nobody offers the second coffee, nobody suggests the starter, nobody reads the table and proposes a dessert to share. That gap doesn't close with hope or a prettier photo: it closes with menu architecture.

Why does average ticket fall if the customer is the same person?

Bundles built on calculated margin, add-ons priced as anchors —a sauce at 4,000 pesos that makes a 38,000-peso combo look reasonable—, order minimums that push the second dish.

An operator who lifts the ticket from 32,000 to 36,000 pesos recovers twelve of the thirty points the channel cost, without selling a single extra order. Settle this before raising prices blind: architecture first, price second. A dish that leaves the pass in four minutes and arrives perfect can show up wrecked at twenty-eight minutes, and that time gap destroys more reputation than any commission. Fries, breaded fried items and pastas with emulsified sauces are the three profiles that travel worst: they lose texture, release fat and arrive at the wrong temperature. Diego F. Parra and the Masterestaurant team run Rappi delivery strategy as a SEPARATE business unit, with its own profit and loss statement, not as an extension of the physical restaurant, and that accounting decision forces the following: the channel menu is designed by transport resistance, not by dining-room popularity.

The delivery menu is not the dining-room menu with a different photo

Stews, rice dishes, bowls, meats in sauce and cold desserts hold up. Cut the three worst travelers from your digital menu this month and measure quality complaints at thirty days. The Medellín operator who showed me 41 million pesos sold on Rappi in one month was missing a line on his dashboard: the two daily hours a cook spent doing nothing but dispatching delivery during the seven o'clock rush. Taken down to contribution margin, after commission, packaging and that station time, less than three million remained out of the 41. The dining room, on half that revenue, kept four times more. Ask yourself what happens if the aggregator raises your commission three points tomorrow: with a channel margin already at 5% or 6%, those three points push you into loss while your dashboard keeps glowing green, because the dashboard measures REVENUE. A separate profit and loss statement isn't accounting vanity, then; it's the only instrument that warns you in time.

Direct ordering is the exit, and 58% of your customers already prefer it

Fifty-eight percent of customers prefer ordering through the restaurant's own app or website rather than an aggregator, according to NCR Voyix as reported by Restaurant Dive in 2024, and that figure dismantles the argument that people 'are already on Rappi and won't move'. People move when you give them a reason and a route. A magnet in the bag with the direct-order WhatsApp number, a 10% discount that still runs twenty points cheaper than the commission, a code printed on the dining-room check. I got this wrong for years, recommending a clean break from the aggregator; it doesn't work, because the aggregator is discovery and the direct channel is repeat purchase. The tension resolves like this: Rappi captures new customers, your own channel keeps them. Set a mix target —25% of total revenue on aggregators, maximum— and chase it quarter by quarter.

The sector's big money is not financing your margin

Agrifoodtech accounts for just 5.5% of global venture capital dollars (AgFunder News, 2024), while the cloud kitchen market is projected at USD 203.72 billion by 2033 (Grand View Research) and ghost kitchens at USD 142.5 billion by 2029 (Research and Markets). Spain alone projects USD 1.379 billion in ghost kitchens by 2032 (Expert Market Research, 2024). I bring this up not so you open a dark kitchen, but so you see who in this chain is capitalized and who isn't: capital funds the channel's infrastructure, not your kitchen. Delivery robots project USD 3.2365 billion by 2030 at a 32.4% CAGR (MarketsandMarkets) and robotic kitchens USD 7.62 billion (Market.us). Translation for your operation: the commission is not going to drop on its own. Your margin has to come out of your menu and your price. First: 30% maximum commission, the ceiling of the most expensive aggregator tier (CloudKitchens Blog 2024).

The 3 numbers you should tattoo on yourself

Action: if your channel price isn't running 20% to 25% above your dining-room price, raise it this week and measure rejection at fourteen days. Second: 58% of customers prefer direct ordering (NCR Voyix via Restaurant Dive, 2024). Action: put a call to your own channel in every bag that leaves through the aggregator for sixty days and count how many come back via WhatsApp. Third: 29.2% user penetration in meal delivery in 2026 (Statista 2026). Action: review your revenue mix and set the 25% aggregator cap; if you sit at 40% today, you have a dependency problem, not a marketing one. Open a spreadsheet tomorrow with the channel's separate profit and loss statement and don't close it until you hold three months of real contribution margin. Commission is not the channel's only cost, merely the visible one. Add packaging, waste from cancelled orders, the courtesy refires you pay when a complaint lands, and the opportunity cost of a station tied up at peak: real leakage usually runs 6 to 11 points above the nominal commission.

Where the money leaks that the aggregator dashboard never shows?

Average ticket on aggregators lands 8% to 12% below the dining room because no server suggests anything. Nobody offers the second coffee or the starter.

You close that gap with menu architecture — bundles, anchored add-ons, order minimums — not with optimism. A dish that plates in four minutes and arrives perfect can arrive ruined at 28 minutes. Fries, breaded fried items and pastas in emulsified sauces are the three usual suspects; publishing them without redesigning packaging buys you two-star reviews that later cost you ranking. Visibility inside the app is bought with discount, and that discount comes out of your margin, not the aggregator's. A 20% discount on a dish carrying 52 points of contribution margin leaves 32: the promotion has to more than double order count just to break even. Once the channel passes 40% of revenue, you no longer own the customer. No phone number, no email, no control over display price, and the aggregator can rewrite commercial terms without asking you. That dependence is a structural risk, not an operating detail.