Rappi delivery strategy: the numbers before and the numbers after
A Rappi delivery strategy only pays when you measure margin BY CHANNEL instead of looking at the restaurant as one block: with aggregator commissions running from 18% to 30% of order value, a dish with 32% food cost that earns money in the dining room can close at a loss inside the app. What moves the needle is not paid placement or discounts, it is rebuilding the digital menu around 12 to 18 low food cost references, pricing the channel 12% to 18% above the dining room, and driving prep time below 18 minutes, which is the variable the algorithm uses to decide who gets shown first.
An owner opens the monthly Rappi report, sees a large gross number and relaxes. Ask what that channel left after commission, packaging, the late courier and the three refunded orders that arrived cold, and almost nobody has the figure. That gap, not the commission, is what kills good restaurants.
At Masterestaurant we have spent twenty years entering through the cash register before the kitchen, and in the digital channel the pattern repeats with boring precision: the aggregator is not a profitability problem, it is an AMPLIFIER of the profitability you already had. A dish that leaves 18 points of contribution margin in the dining room leaves four or five in the app. One that left 65 points still runs a business.
The figures below come from public sector sources —National Restaurant Association, Statista, Technomic, the platforms' own quarterly reports— and they are grouped by the decision they trigger, not by topic. At the end you get the three worth memorizing, each with the action it demands.
One warning first. None of this works if your digital menu is the dining room menu photographed and uploaded. That mistake looks cosmetic and explains most delivery operations that bill a lot and keep nothing.
Side-by-side comparison
| Before: Rappi with no channel strategy | After: Rappi with the Masterestaurant method | |
|---|---|---|
| Effective commission per order | ✕22% to 30% paid on dining room prices, no list adjustment | ✓22% to 30% absorbed with channel price +14%, 9 margin points recovered |
| References on the digital menu | ✕48 to 90 dishes copied from the physical menu, 34% average food cost | ✓12 to 18 curated references, 26% average food cost |
| Stated preparation time | ✕25 to 34 real minutes, 11 minutes above the promise | ✓14 to 18 minutes with a dispatch station separated from the line |
| Channel contribution margin | ✕3 to 6 points, discovered at year end by surprise | ✓17 to 23 points, reviewed every Monday on a channel P&L |
| In-app rating | ✕4.1 to 4.3 stars with 6% of orders carrying an incident | ✓4.7 to 4.9 stars with incidents under 2% |
| Channel average ticket | ✕USD 11, no combos or upsells, 1.7 items per order | ✓USD 16 with 3 combos and a drink prompt, 2.6 items per order |
| Aggregator dependency | ✕94% of delivery volume inside the app, zero owned customer data | ✓61% aggregator and 39% owned channel via WhatsApp and Google Business Profile |
Why your restaurant-wide margin lies to you about Rappi?
A consolidated margin lies because it averages a channel that pays 18% to 30% in commission with one that pays nothing, and that average hides exactly where the profit is leaking.
Statista projects worldwide online food delivery at USD 1.51 trillion for 2026, growing at a 6.24% CAGR through 2031, and the United States alone at USD 473.49 billion this year; none of those figures helps you if your books have a single column. Run it with cash numbers: a USD 10 dish at 32% food cost leaves USD 6.80 of gross margin in the dining room; that same dish on the app, with 25% commission and USD 0.45 of packaging, leaves USD 3.85. Still positive, but contribution dropped by 43 percentage points and you never saw it, because the P&L added it all together. Rappi and its peers move volumes that justify being there, and that is precisely the argument that bleeds restaurants which never split their accounts.
Channel size is not the question; how much is left is
Uber Eats reported USD 74.6 billion in gross bookings for 2024 in its Form 8-K filed with the SEC; DiDi Food Mexico declared roughly 74,000 restaurants on its app, 70% of them local small businesses (DiDi Food, 2024); Swiggy closed fiscal 2023-24 with 196,000 partner restaurants across 653 cities serving some 13 million users. Those three data points say the same thing: demand exists and it is not going back to the dining room. But volume and profit are independent variables. A channel can double your revenue while draining your cash flow, and we have watched operations grow sales 60% and drop their EBITDA. Block takeaway: enter the channel, yes, but enter with its own P&L from the very first order. Split the books into three columns and the Rappi conversation stops being emotional and turns into arithmetic. At Masterestaurant, Diego F. Parra insists on entering through the cash register before the kitchen, and that spreadsheet is the first thing we ask for: sales, food cost, packaging, commission and refunds, channel by channel, month by month.
Three income statements: dining room, aggregator, own channel
The dining-room column carries rent and floor payroll; the aggregator column carries 18% to 30% commission, packaging and the waste from rejected product; the own-channel column carries ad spend and payment gateway, which rarely exceeds 4%. Once those three columns exist, the decision makes itself: you keep on the app the dishes whose contribution margin survives 25 points of haircut, and you pull the ones that do not. Some 76% of US operators believe technology gives them a competitive edge (National Restaurant Association, 2024), but the edge is in measuring, not in showing up. Charging more on the app is legitimate, it is standard practice, and the guest accepts it, because whoever orders delivery is buying convenience rather than price. They already pay the delivery fee, they tip, and they pay for not leaving the house; a 12% to 15% uplift over the dine-in price sits inside what they consider fair.
Channel pricing: the boardroom panic your guest never notices
What they never forgive is a cold plate. Do the math: on that USD 10 dish, a 14% adjustment adds USD 1.40 that lands whole against the commission, and channel contribution climbs from USD 3.85 to USD 5.25 without touching the recipe or the supplier. The standard boardroom objection is reputational risk from price comparison. It has gone twenty years without showing up in any measurable volume, while the cost of not adjusting shows up in the cash flow every single month. Photographing your dine-in menu and uploading it as is explains most delivery operations with heavy revenue and zero profit. A risotto that comes out perfect three meters from the pass arrives gummy after thirty minutes, and the bleeding starts right there: full order refund, food cost gone, packaging gone, commission the platform sometimes never returns. Three refunds out of a hundred USD 10 orders erase USD 30 of sales, yet the real cost lands near USD 44 once you add the ingredients and packaging on those refunded orders.
The cloned menu mistake and the waste nobody books
In the dining room that mistake gets fixed at the table: the server takes the plate back, the kitchen fires another, the guest leaves happy. On the app there is no second chance, there is a one-star rating. Block takeaway: audit your digital menu dish by dish and drop, without nostalgia, anything that cannot survive thirty minutes in a box. Delivery's cost structure will change before 2030, and whoever measures by channel will get to decide with data when to move. Global Growth Insights projects the worldwide dark kitchen market at USD 171.3 billion by 2033; Grand View Research puts food robotics at USD 6.81 billion by 2030, at a 20.6% CAGR; MarketsandMarkets estimates delivery robots at USD 3.2365 billion by 2030, compounding at 32.4% a year. In Spain, quick commerce would reach USD 4.37 billion by 2029 according to Research and Markets.
Dark kitchens and automation: where the channel's cost is heading
Translated into your operation: a ghost kitchen strips out floor rent and dining-room payroll, so a dish that cannot carry Rappi's commission today may well carry it inside a structure with no dining room. Block takeaway: do not chase the technology, chase the channel's break-even point, and let the technology in when it moves that number. Something uncomfortable happens, and then something good: channel average ticket drops for about six weeks, and contribution margin rises from the first cut. The panic around pulling the signature dish assumes the digital guest is the same one who sits down, and they are not. App buyers pick formats that travel, and your signature dish is usually the one that travels worst and generates the most refunds. Assume a hundred monthly orders of it at USD 10, 25% commission and an 8% refund rate: that is USD 1,000 in sales leaving roughly USD 300 of net contribution.
What happens if you pull your signature dishes off the app?
Replace it with two delivery-engineered formats contributing USD 5.50 each, and 55 orders match it. I got this wrong for years, recommending full menu presence on digital.
The right criterion is not what sells most, it is what survives thirty minutes on a motorcycle. Three figures with their concrete action, no ornament. FIRST: aggregator commission runs from 18% to 30% of order value, and that band defines everything else. Action: today, open a sheet with three columns —dining room, aggregator, own channel— and load last closed month's sales into it. SECOND: online food delivery will move USD 1.51 trillion in 2026 at a 6.24% CAGR through 2031 (Statista, 2026), which means the channel is not a fad you can outwait. Action: set digital menu prices 12% to 15% above dine-in and measure it for six weeks against volume. THIRD: 76% of US operators consider technology a competitive advantage (National Restaurant Association, 2024), though the advantage sits in the data, not in the tool.
The 3 numbers worth tattooing
Action: pull from the app every dish whose per-channel contribution margin falls below 20% after commission, packaging and refunds. The difference is not technological, it is accounting. A restaurant with no channel strategy runs ONE income statement; one with a strategy runs three —dining room, aggregator, owned channel— and decides differently in each. The day you split those columns, the Rappi conversation stops being emotional and becomes arithmetic. Channel pricing terrifies the boardroom and then turns out to be the least contested change. Delivery customers buy convenience, not price: they already pay the fee, the tip and the comfort of staying home. Fourteen percent more for the same dish reads as reasonable; what they never forgive is food arriving cold. In the dining room a mistake gets fixed at the table: the server takes the plate back, the kitchen replaces it, the guest leaves happy. In the app there is no second chance —the error becomes a star, the star becomes less visibility, and less visibility becomes fewer orders next week.
What actually changes between the two operations?
A badly dispatched delivery order costs four to seven times the cost of the dish. A dark kitchen or ghost kitchen solves occupancy cost and does nothing for acquisition cost:
without a dining room you lose the walk-in guest and depend entirely on the algorithm. That is the paradox of the channel, and it resolves with one rule: a cocina oculta makes sense once you have proven demand for a brand, never as a way to discover it. The aggregator hands you volume and takes your customer. You will never know the phone number of the person who ordered forty times this year, and that asymmetry is the real price of the channel, dearer than the commission. Which is why the metric I watch hardest is not Rappi sales, it is the share of delivery that no longer goes through Rappi.
Criterion by criterion
Rappi with no channel strategyWhere almost everyone starts
- The digital menu duplicates the physical one, same prices and the same 60 references.
- Margin gets reviewed once a year, when the accountant closes a consolidated P&L.
- The kitchen fires delivery orders between dining room tickets, with no station and no sequence.
- Packaging is bought on price, never on how the dish behaves thermally inside it.
- When sales dip, the answer is in-app advertising or a 30% discount funded by the house.
- The Google Business Profile listing sits outdated and local search brings in no direct orders.
Rappi with the Masterestaurant methodMasterestaurant
- A channel menu built from scratch with references that travel well and hold food cost under 28%.
- Channel list price 12% to 18% above the dining room, with commission modeled inside the price.
- A separate dispatch station with its own sequence and a named owner on every shift.
- Channel P&L closed every Monday: commission, packaging, refund waste and allocated kitchen hours.
- Five-star reviews worked on Rappi and on Maps under the same protocol, because they feed different algorithms.
- An owned channel running in parallel: WhatsApp Business, ordering from the Maps listing and a QR menu on the house database.
Side-by-side comparison
| Before: Rappi with no channel strategy | After: Rappi with the Masterestaurant method | |
|---|---|---|
| Effective commission per order | ✕22% to 30% paid on dining room prices, no list adjustment | ✓22% to 30% absorbed with channel price +14%, 9 margin points recovered |
| References on the digital menu | ✕48 to 90 dishes copied from the physical menu, 34% average food cost | ✓12 to 18 curated references, 26% average food cost |
| Stated preparation time | ✕25 to 34 real minutes, 11 minutes above the promise | ✓14 to 18 minutes with a dispatch station separated from the line |
| Channel contribution margin | ✕3 to 6 points, discovered at year end by surprise | ✓17 to 23 points, reviewed every Monday on a channel P&L |
| In-app rating | ✕4.1 to 4.3 stars with 6% of orders carrying an incident | ✓4.7 to 4.9 stars with incidents under 2% |
| Channel average ticket | ✕USD 11, no combos or upsells, 1.7 items per order | ✓USD 16 with 3 combos and a drink prompt, 2.6 items per order |
| Aggregator dependency | ✕94% of delivery volume inside the app, zero owned customer data | ✓61% aggregator and 39% owned channel via WhatsApp and Google Business Profile |
Channel numbers, grouped by the decision they trigger
“We were billing 96 million pesos a month on Rappi and we were sure it was our best channel. Once we built a separate P&L, that channel left 4 points of margin while the dining room left 22. We cut the digital menu from 71 dishes to 15, raised channel prices 15% and built a dispatch station with its own shift owner. Four months later revenue fell to 88 million and channel margin climbed to 19 points: we gave up eight million in sales and gained fourteen in profit.”
Rebuilding your delivery strategy in four steps
Pull the platform settlement report and subtract, in this order, commission, packaging, refunds, promotions you funded and the kitchen hours spent firing delivery. Divide what remains by channel gross sales and you have your real margin, which will land 10 to 18 points below the dining room. That number is the starting point for everything else; without it, any Rappi decision is an expensive hunch.
Rank dishes by contribution margin in currency, not percentage, and keep the ones that survive thirty minutes inside a box. Fried items that go soft, salads that sweat and sauces that split come off the digital menu even when they star in the dining room. Twelve to eighteen references are enough: an app customer decides in forty seconds, and a long menu lowers conversion while raising prep time.
Take the dining room price, divide by one minus the effective commission and round up: at 25% commission a USD 10 dish needs USD 13.3 to leave the same money. Raise between 12% and 18% —above that customers do notice— and cover the rest with combos that lift the ticket. Transparency is non-negotiable here: if a guest asks, the answer is that app pricing includes the cost of the channel.
Build a station with its own mise en place, its own shift owner and its own assembly sequence, because a delivery order competing with dining room tickets always loses. Time twenty real orders rather than trusting the POS, then hunt the bottleneck: it usually sits in packaging, not in cooking. Once you break eighteen minutes the algorithm starts ranking you higher and sales rise without a cent of advertising.
And with AI?
Optimize channels, pricing and unit economics of your dark kitchen. Diego F. Parra is an expert in AI applied to restaurants.
Free tools to apply this now
Ecosystem tools that hold the decision together
These three tools solve the three bottlenecks of this work: the business model of the channel, growth of your own demand, and cash control during the transition. None of them replaces the channel P&L; they make it sustainable.
Diego F. Parra — International consultant, expert in creating and scaling restaurants and in AI applied to restaurants, foodtech and HORECA. Methodology applied in 8.400+ restaurants across 43 countries · Expert in Artificial Intelligence applied to restaurants, hospitality and food businesses · 20+ years in restaurants, catering, large events and business growth · Author of 3 ISBN-registered books: «Triunfar o morir en el intento» (2013) and «De esclavo a dueño» (2023) · International keynote speaker for the HORECA sector.
Questions owners ask me about this channel
How much commission does Rappi charge a restaurant in 2026?
How much commission does Rappi charge a restaurant in 2026?
Delivery aggregator commissions run between 18% and 30% of order value depending on country, contracted plan and visibility tier, and Technomic 2026 places the ceiling at 30% for maximum-exposure plans. On top of that you still carry packaging, refunds and co-funded promotions, so the real channel cost usually lands between 28% and 38%.
Should I open a dark kitchen instead of selling from my location?
Should I open a dark kitchen instead of selling from my location?
Only if you already have proven demand for that brand and your bottleneck is capacity rather than discovery. A ghost kitchen removes dining room occupancy cost, and it also removes the walk-in guest, leaving you dependent on the aggregator algorithm to exist. In dark kitchen vs brick and mortar restaurant the criterion is simple: a cocina oculta scales demand that already exists, it never creates it.
Is it better to raise app prices or absorb the commission?
Is it better to raise app prices or absorb the commission?
Raise prices, 12% to 18% above the dining room. Absorbing commission out of house margin turns every incremental order into a bigger loss, which is the classic trap of the channel. Delivery customers buy convenience and accept the differential without friction; what they will not tolerate is cold food or an incomplete order, per National Restaurant Association 2026.
Should I drop the physical menu now that I have a QR menu and Rappi sales?
Should I drop the physical menu now that I have a QR menu and Rappi sales?
No. The physical menu stays, always, because it is the instrument you use to control service pace, menu narrative and suggestive selling at the table. The QR is a complement: it serves delivery, accessibility, price changes and analytics on what guests browse. Each has its role, and removing the physical one costs you average ticket in the dining room.
Sector data 2026 (official sources)
Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.
| Metric | Benchmark 2026 | Source |
|---|---|---|
| Mercado de ghost/cloud kitchens | mercado global en fuerte crecimiento de doble dígito (CAGR) | Statista · Ghost kitchens |
| Estructura de la industria de ghost kitchens (EE.UU.) | tamaño y número de operaciones en informe de industria | IBISWorld · Ghost Kitchens (US) |
| Mercado global cloud/ghost kitchen 2026 | USD 88.7 mil millones en 2026; CAGR 12.6% (2026-2033) | Grand View Research 2026 |
| Mercado cloud kitchen 2026 (proyección alterna) | USD 83.5 mil millones en 2026; CAGR 9.7% al 2034 | Fortune Business Insights 2026 |
| Cloud kitchen al 2035 | USD 248.10 mil millones proyectados para 2035 | Precedence Research 2025 |
| Reparto de comida en línea mundial 2026 | USD 1.51 billones en 2026; CAGR 6.24% (2026-2031) | Statista 2026 |
Related content
Grow your restaurant with the Masterestaurant method
Applied in +8.400 restaurants across 43 countries.