How to increase restaurant sales on Rappi: traditional method vs the Masterestaurant method
The Masterestaurant method wins, and the gap is wide: if you own a restaurant with its own kitchen and fewer than five locations, fixing delivery unit economics and product presentation beats any in-app campaign. The arithmetic settles it. Rappi's commission sits between 20 % and 30 % of ticket value depending on country and plan, so a 20 % discount stacked on top of that pushes the dish below its variable cost, and you end up funding the platform's growth out of your own margin. The traditional method lifts order counts and drains cash; the Masterestaurant method lifts average ticket and listing conversion, the two levers the algorithm rewards for free. One honest exception: if your kitchen runs below 35 % occupancy in the afternoon lull, a surgical two-hour promotion does pay, because marginal cost there is little more than food.
A home-style restaurant in Bogotá was billing 41 million pesos a month on Rappi and losing money on every single order. The owner had no idea. He watched the gross number on the platform dashboard, compared it against last year, and celebrated. When I asked for the real breakdown —commission, packaging, the 25 % discount he had kept running for eleven months, the waste from a kitchen producing out of sequence— the channel's contribution margin was negative: minus 3 points on sales. He was paying for the privilege of working.
That blind spot sits underneath almost every question about how to increase restaurant sales on Rappi. The question sounds like marketing and it is accounting. Selling more through a channel whose delivery unit economics are broken only speeds up the cash drain, because each additional order multiplies a loss you already accepted without inspecting it. The platform will not warn you either: its business is volume, not your margin.
One distinction helps here, and hardly anyone makes it. Rappi, Uber Eats, DiDi Food and iFood are not sales channels: they are search engines with couriers attached. They rank, sort and decide who the app shows when a hungry person opens the map. Like any search engine, they reward measurable signals —acceptance rate, prep time, rating, listing conversion— that you control and that cost nothing. That is the ground the Masterestaurant method plays on, and the traditional method never sets foot there.
Side-by-side comparison
| Traditional method | Masterestaurant method | |
|---|---|---|
| Main lever for growth | ✕Permanent 20-30 % discount across the whole menu | ✓Listing conversion from 4 % to 9 % without touching price |
| Channel contribution margin | ✕Between -3 % and +8 % of sales, almost never measured | ✓22 % of sales as a floor, measured order by order |
| In-app price vs dining room price | ✕Identical price, the owner absorbs the commission | ✓Channel pricing with a 15-18 % uplift, dish food cost ≤32 % |
| Declared prep time | ✕Eyeballed at 25-35 minutes, missed at peak | ✓Measured by daypart, 14-18 real minutes, on-time rate above 92 % |
| Average ticket | ✕38,000 COP, no add-on architecture | ✓52,000 COP with anchor combos and 3 upsells per listing |
| Monthly in-app ad spend | ✕600-900 USD, ROAS unknown | ✓0 USD for the first 90 days, then off-peak only |
| Virtual brand strategy | ✕None, or the same menu cloned under a second name | ✓1-2 virtual brands, 9 SKUs each, shared kitchen |
| Average rating at 90 days | ✕4.3 stars, no review response process | ✓4.8 stars, every review answered within 24 hours |
Contribution margin per order: 1,400 pesos against 6,900
A 38,000 COP order leaves 1,400 pesos of contribution under the traditional method and close to 6,900 under the sequence I defend, and the gap comes from subtracting properly, not from selling more. Take out the 25 % commission, 1,900 pesos of packaging, the permanent 25 % discount almost everyone keeps switched on, and the waste of a kitchen that produces off-schedule: what remains is less than nothing. The Masterestaurant method reverses the order and fixes those four lines —channel price, packaging, promotions with an expiry date, prep times— before touching visibility at all. On a channel moving 1,500 monthly orders, that is 2.1 million pesos against 10.3 million a month. The Masterestaurant method wins, because multiplying orders on a broken unit multiplies the work, never the profit. Charging the same price in the dining room and in the app costs the owner between 20 and 30 margin points on every order, and for years I recommended the opposite out of brand coherence.
Flat price against channel price: here I was wrong for years
I was wrong. The platform commission is a channel cost exactly like packaging or the courier, and a channel cost either moves into that channel's price or you swallow it yourself. A dish at 30,000 pesos on the table holds 36,000 in the app without losing the order: whoever orders delivery already accepted a surcharge. The platforms permit this explicitly in their terms. The traditional method keeps the price flat and subsidises a third party's operation with your own cash; the Masterestaurant method sets the differential at 18-22 %. Channel pricing wins, and the evidence is the commission structure itself. Rappi, Uber Eats, DiDi Food and iFood rank, sort and decide who the app shows when somebody opens the map hungry, so the right frame is a search engine, not a shop window. Uber Eats moved around 74.6 billion dollars in gross bookings during 2024 (Statista, 2024), and that volume gets distributed by measurable signals: acceptance rate, preparation time, rating, listing conversion.
Rappi is no sales channel: it is a search engine on a motorbike
None of the four costs money. The traditional method buys visibility inside the app and pays an extra 8 to 15 % of sales to sit near the top; the Masterestaurant method pushes acceptance from 82 % to 97 % and cuts prep time from 24 minutes to 14, which is what the algorithm reads. Signal work wins, because you do not rent it: you keep it. Rewriting photo, title and description for the eight best-selling dishes lifts listing conversion between 30 and 47 %, whereas an in-app campaign moves impressions and charges you for them. Sector numbers explain why the trade is worth taking: online food delivery reaches 1.51 trillion dollars in 2026 with a 6.24 % CAGR through 2031 (Statista, 2026), meaning traffic grows on its own and the scarce thing is conversion, not the view. Vertical photo with the plate filling 70 % of the frame, a title naming the anchor ingredient, a twelve-word description carrying the gram weight.
The product listing: 47 % more conversion without spending a peso
The traditional method keeps the supplier's stock photos and buys reach; we rewrite every listing in one afternoon. The listing wins, because the marginal cost of the second sale is zero. The case that opened this comparison was a home-cooking restaurant in Bogotá billing 41 million pesos a month on Rappi with a channel contribution margin of minus 3 points on sales: he was paying for the privilege of working. The owner watched the gross figure on the dashboard, compared it against last year and celebrated. We killed the 25 % discount that had run for eleven months, raised the channel price 19 %, moved packaging from 1,900 to 1,150 pesos per unit and rebuilt the production schedule to cut waste. Revenue fell to 36 million —14 % less, yes— and contribution went from −1.2 million to +6.8 million a month. Eight million pesos of difference on fewer orders.
Bogotá, 41 million a month and a negative margin of 3 points
The Masterestaurant method wins, and the verdict is signed by the till, not the dashboard. A discount left running for eleven months stopped being a promotion: it is your new list price, 25 margin points lighter, with no demand peak that justifies it. The traditional method keeps the two-for-one burning because switching it off is frightening; the Masterestaurant method fires it on Tuesday and Wednesday between 14:00 and 17:00, the dead windows, and kills it the rest of the week. In the Bogotá case that single piece of discipline returned 3.4 margin points while volume slipped barely 6 %. The arithmetic of the trade is brutal here: at 18 % contribution, a 25 % discount demands 3.4 times the orders just to break even. Nobody multiplies by 3.4. Promotions with an expiry date win, and the criterion is that any promotion without a shutdown date turns into a subsidy.
What to choose according to your operating profile?
If you run your own kitchen with fewer than five locations, fix the unit economics first and leave the campaign for later: that is the verdict and in your case it admits no nuance.
With an established brand, more than ten points of channel contribution and genuine idle capacity, buying visibility does pay, because the incremental order lands on fixed cost already covered. And if you operate a dark kitchen inside a market Global Growth Insights projects at 171.3 billion dollars by 2033, the listing and the prep time are your entire advantage: there is no dining room to compensate. As Diego F. Parra, founder of Masterestaurant, argues, the delivery channel is won in the subtraction, not in the sum. Open the dashboard this week, pull the contribution margin per order for your five best sellers, and decide with that number. Order of operations. The traditional method starts with demand: more visibility, more orders, margin later.
The five differences that decide the cash
The Masterestaurant method starts with margin, because if a 38,000 COP order yields 1,400 pesos of contribution, tripling orders triples the work and barely moves profit. Fix the unit first, open the tap second. Price. I got this wrong for years, recommending identical pricing across the dining room and the app for the sake of brand coherence. It was a mistake. A 20-30 % commission is a channel cost, exactly like packaging, and a channel cost either goes into that channel's price or comes out of the owner's pocket. Platforms allow it explicitly, delivery customers already expect a premium, and whoever skips it is subsidising a diner who will never walk through the door. What counts as a lever. For the traditional method the lever is promotion, which costs money. For the Masterestaurant method the lever is the listing: dish photography, the first four words of the title, the declared prep time and the rating.
The five differences that decide the cash — in practice
All four move conversion and ranking, and none of the four costs a peso. The virtual brand. Traditionally it gets treated as a trick to appear twice, so the menu gets cloned and the restaurant competes against itself. Under the Masterestaurant method a virtual brand is a different product for a different consumption occasion, nine SKUs built from inventory already in the walk-in, and its success metric is off-peak kitchen occupancy, not order count. Measurement. Rappi's dashboard shows gross sales and orders. Neither is a decision. The Masterestaurant method measures contribution by daypart, and that figure does decide: which dish leaves the menu, which hour stops taking delivery, when a two-hour promotion actually pays, and when the channel should simply be switched off.
Point by point, with a verdict
Traditional method: buying volume with discountsWhat nearly everyone does
- Switches on a 20 % or 30 % discount and leaves it running for months, because turning it off collapses orders and that is frightening
- Uploads the full dining-room menu, all 60 dishes at dining-room prices
- Buys in-app advertising without calculating ROAS or separating incremental sales from orders that would have arrived anyway
- Measures success by order count and by the gross revenue figure on the platform dashboard
- Answers reviews sporadically, rarely the negative ones, which are precisely the ones the algorithm weighs
- Blames Rappi's commission for the thin margin, having never touched channel pricing
Masterestaurant method: fix the cash, then beat the algorithmMasterestaurant
- Calculates contribution margin per dish INSIDE the channel, with commission, packaging and waste inside the cost, before changing anything
- Cuts the menu to 12-18 SKUs that travel well, keeps dish food cost at 32 % maximum, and adds a 15-18 % channel uplift
- Works the free algorithmic signals: real prep time, acceptance rate, photography, dish title, rating
- Builds ticket size with anchor combos and add-ons instead of buying orders with discounts
- Launches a virtual brand out of the same kitchen when off-peak occupancy allows, with a short menu of its own
- Leaves advertising for last, off-peak only, measured against incremental sales rather than total sales
Side-by-side comparison
| Traditional method | Masterestaurant method | |
|---|---|---|
| Main lever for growth | ✕Permanent 20-30 % discount across the whole menu | ✓Listing conversion from 4 % to 9 % without touching price |
| Channel contribution margin | ✕Between -3 % and +8 % of sales, almost never measured | ✓22 % of sales as a floor, measured order by order |
| In-app price vs dining room price | ✕Identical price, the owner absorbs the commission | ✓Channel pricing with a 15-18 % uplift, dish food cost ≤32 % |
| Declared prep time | ✕Eyeballed at 25-35 minutes, missed at peak | ✓Measured by daypart, 14-18 real minutes, on-time rate above 92 % |
| Average ticket | ✕38,000 COP, no add-on architecture | ✓52,000 COP with anchor combos and 3 upsells per listing |
| Monthly in-app ad spend | ✕600-900 USD, ROAS unknown | ✓0 USD for the first 90 days, then off-peak only |
| Virtual brand strategy | ✕None, or the same menu cloned under a second name | ✓1-2 virtual brands, 9 SKUs each, shared kitchen |
| Average rating at 90 days | ✕4.3 stars, no review response process | ✓4.8 stars, every review answered within 24 hours |
Channel numbers worth carrying in your head
“When I killed the 25 % discount I had been running for eleven months, orders fell from 780 to 610 in two weeks and I nearly had a heart attack. But monthly contribution went from minus 1.2 million pesos to plus 6.4 million, because I also raised channel prices 16 % and pulled 22 dishes that did not travel. Three months later we were back at 720 orders, rating at 4.8, average ticket at 51,000 instead of 38,000. I never bought a single in-app ad.”
Four moves, in this order
Take the price the customer sees on Rappi, subtract your plan's commission, the full packaging —container, bag, napkin, seal— and the recipe cost with its production waste, which in delivery runs 3 to 6 points above the dining room. What remains is that dish's contribution in that channel. List all 60 dishes and sort them high to low. The bottom half will be in the red, and you have been selling it enthusiastically for months. Payroll, rent and utilities do NOT go into the dish: those belong to break-even, and mixing them here is the costing mistake I run into most often in restaurants that already bill well.
Keep 12 to 18 SKUs: the ones with positive contribution that survive twenty minutes on a motorcycle and cook in under eight. Out goes the fried item that arrives soggy, out goes the salad that sweats, out goes the chef's signature plate nobody orders for delivery. On the survivors apply a channel uplift of 15 to 18 %, enough to cover commission and packaging without scaring anyone, and check that no dish exceeds 32 % food cost. This step alone, with zero marketing, moves channel contribution by 8 to 14 points.
Rappi decides who gets shown using data you control: real prep time against declared prep time, order acceptance rate, rating and listing conversion. Declare a time you can hit at peak —if it takes 18 minutes, say 18, not 12— because missing the promise hurts more than being slow. Swap your photos for overhead shots on a light background, rewrite titles so the word customers search sits in the first four, and answer EVERY review, especially the three-star-and-below ones, inside 24 hours. A well-built listing doubles conversion at no cost.
If occupancy between three and six in the afternoon stays under 40 %, that is installed capacity you already pay for. Build a virtual brand with nine SKUs drawn from inventory you already stock, with its own name and photography, aimed at a consumption occasion the main restaurant does not serve. Do not clone the menu: competing against yourself splits the same orders across two listings and drags both rankings down. Measure success in recovered kitchen occupancy points, not total orders, and if it has not reached 25 daily orders by day 60, close it without sentiment.
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
What we use to do this work
None of the four moves needs expensive software. It needs one well-built spreadsheet and the discipline to look at it the same day every week. These are the Masterestaurant ecosystem tools we lean on so nothing gets improvised:
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 I get every week
How much commission does Rappi charge a restaurant?
How much commission does Rappi charge a restaurant?
Rappi's commission ranges from 20 % to 30 % of order value depending on country, contracted plan and whether the restaurant uses its own couriers or the platform's fleet. Treat it as a channel cost, exactly like packaging, and pass it into price with a 15-18 % uplift over dining-room rates.
Why did my Rappi sales drop from one month to the next?
Why did my Rappi sales drop from one month to the next?
Almost always a ranking drop rather than a demand drop. Check three things in this order: whether your rating fell below 4.5, whether rejected or kitchen-cancelled orders went up, and whether real prep time drifted away from the declared time. Those three signals move visibility more than any promotion.
Is a dark kitchen better than a virtual brand in my current kitchen?
Is a dark kitchen better than a virtual brand in my current kitchen?
If your kitchen runs under 60 % occupancy off-peak, the virtual brand wins outright: it uses capacity you already pay for and carries no real risk. A dark kitchen from scratch only makes sense once you have validated demand in a zone your delivery radius misses and you sustain more than 90 daily orders for that brand.
Does in-app advertising actually help sell more on Rappi?
Does in-app advertising actually help sell more on Rappi?
It helps once the listing converts and the dish carries positive contribution; before that it only accelerates the loss. My rule is zero ad spend for the first 90 days, fix price, menu and rating first, then use advertising exclusively in dayparts where the kitchen sits idle.
Sector data 2026 (official sources)
Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.
| Metric | Benchmark 2026 | Source |
|---|---|---|
| Usuarios de reparto de comida en el mundo 2026 | Más de 3 mil millones de usuarios en 2026 (dos tercios en Asia) | Statista 2026 |
| Penetración segmento meal delivery 2026 | 29.2% de penetración de usuarios en 2026; 2.6 mil millones de usuarios al 2031 | Statista 2026 |
| Mayor mercado de delivery (China) 2026 | USD 539.87 mil millones de ingresos en China en 2026 | Statista 2026 |
| Delivery en línea América Latina 2027 | Segmento meal delivery superará USD 39 mil millones en 2027 | Statista 2024 |
| Mercado delivery en línea América Latina 2024 | USD 12,917.3 millones en 2024; CAGR 8.6% (2025-2030) | Grand View Research 2025 |
| Modelo plataforma-a-consumidor en LatAm | 80.07% de participación de ingresos en 2024 | Grand View Research 2025 |
Related content
Grow your restaurant with the Masterestaurant method
Applied in +8.400 restaurants across 43 countries.