How to increase restaurant sales on Rappi: traditional method vs Masterestaurant method

How do you grow Rappi sales without giving up the margin?

You grow them by costing the Rappi dish as a product separate from the dining room, with commission inside the cost instead of outside it.

The traditional method publishes the dining-room menu as is, accepts the 30% discount the aggregator suggests and trusts volume to make up the difference; on a dish costed at 32% food cost, that discount wipes out the entire contribution. The Masterestaurant method builds a short digital menu with its own price, and commission —which in Colombia runs between 22% and 28% depending on the plan— goes into the item cost alongside packaging and transit shrink, which on hot dishes with sauce runs 3% to 6%. The second approach wins for an arithmetic reason: the channel grows on measured contribution, not on counted orders. Online delivery in Latin America moved USD 12,917.3 million in 2024 and grows 8.6% a year through 2030, according to Grand View Research, so the channel is worth it; badly costed, it is not.

The Medellín grill: 41 million on Rappi and 4% of the profit

A 140-seat grill in Medellín was billing 41 million pesos a month on Rappi and the owner celebrated that number, until we split the P&L by channel and the hole showed up: delivery brought 34% of sales and 4% of the profit. The math was simple once you looked at all of it: a 26% commission, an average 22% discount and 1,900 pesos of packaging per order ate whatever the dining room earned the hard way. Nobody had lied to that owner. Nobody had costed anything either. Under the traditional method he would have raised the promotion budget to «win back ranking»; under ours we pulled 31 of the 54 published items, raised the digital price 11% on the six dishes that could carry the commission and cut the discount from 22% to 9%. Four months later he billed 36 million, five million less, and delivered 19% of the business profit.

Long menu versus short menu: where digital margin leaks

The short digital menu wins every time, and not for aesthetic reasons: a 60-item menu multiplies inventory, waste and line time without moving the ticket. According to Aaron Allen, founder of Aaron Allen & Associates, oversized menus destroy margin exactly there, in the hidden cost of keeping references nobody orders. On Rappi the damage doubles, because prep time feeds the visibility algorithm: every odd item that takes 14 minutes to cook punishes the ranking of the 12 dishes that leave the pass in 6. The rule we apply at Masterestaurant is 12 to 18 items in the digital channel, grouped in three fast-decision blocks, and none above 32% food cost after packaging. An owner with 54 published references is paying for visibility to sell confusion, while 80.07% of LatAm delivery revenue flows through the platform-to-consumer model he does not control, per Grand View Research 2025. Differentiated digital pricing beats aggressive discounting in any scenario you follow to the end.

A 30% discount against a differentiated digital price

Take a 30,000-peso dish with 30% food cost, 1,900 pesos of packaging and a 26% commission: it leaves 5,300 pesos of contribution, not the 21,000 that show up on the dining-room spec sheet. Now apply the suggested 30% discount: price drops to 21,000, commission falls to 5,460 and contribution turns NEGATIVE by 340 pesos per unit. What happens if that dish becomes the app's best seller? It sells 600 units a month, gains 18 ranking points, the owner celebrates the growth and loses 204,000 pesos, plus the opportunity cost of a kitchen tied up. The opposite road —raising the digital price 12% to 18% over the dining room, which is what the chains have done since 2022— holds the 5,300 pesos and funds the commission without touching the dining room. It hurts less and lasts longer. Contribution per order is the only metric worth deciding on inside an aggregator, and the app dashboards never show it.

Order volume against contribution per order: the metric that decides

Eight years of platform panels trained the industry to watch orders, GMV and rating; none of those three numbers tells you whether the channel pays payroll. Diego F. Parra insists on a two-line calculation every week in Masterestaurant diagnostics: channel net sales minus commission, discounts, packaging and shrink, divided by order count. If the result falls under 6,000 pesos per order on a 30,000-peso average ticket, the dining room is financing the channel. Here sits the paradox almost nobody resolves: the aggregator DOES work for growth, since off-premise operation already accounts for roughly 75% of industry traffic according to Circana, but it works as a positive-contribution channel, never as a cheap volume channel. Growing at 4% profit is not growth, it is renting out your kitchen. I got this wrong for years: I used to recommend entering the aggregators with the full dining-room menu so as «not to lose sales», and it was exactly backwards.

Packaging and shrink: the 1,900 pesos nobody puts in the spec sheet

The twin mistake, just as expensive, is leaving packaging out of the item cost because «it goes into overhead». Decent packaging for a hot dish with sauce costs between 1,400 and 2,200 pesos in Colombia today, plus lid, cutlery and bag; against a 30,000-peso ticket that is 4.6 to 7.3 points of food cost the traditional method simply does not see. Add 3% to 6% transit shrink on those same dishes, plus the resends from complaints, which reach 2% of orders in operations without a sealing protocol. The Masterestaurant method puts all three line items inside the digital item cost before setting price, and that single accounting decision changes which dishes deserve to be published. Your own promotion wins because it builds an asset, while the aggregator's rents traffic that switches off the day you stop paying. A combo at 22% off inside the app lifts orders while it runs and returns demand to the prior level within two weeks, margin already spent.

Aggregator promotion against your own: who owns the customer

A combo designed at 26% food cost after packaging, pushed over WhatsApp to your own list and delivered by your own driver at 4,500 pesos a trip, leaves 9,000 to 11,000 pesos of contribution and hands back the customer data that inside the aggregator is never yours. More than 25% of operators already use artificial intelligence for tasks of this kind, according to the National Restaurant Association, and that is where the concrete opportunity sits: combo recommendation and demand forecasting by time slot. Use Rappi to capture the stranger; use your own list to bring them back. That order matters more than any ranking tactic. If delivery weighs under 15% of sales and you have no P&L by channel, leave pricing alone for now: measure contribution per order for four weeks and cut the digital menu to 15 items, which is the cheapest intervention and the one that moves the most.

What to choose based on your restaurant profile?

If delivery sits between 15% and 35%, the Masterestaurant method is the only defensible one: digital price 12% to 18% above the dining room, commission and packaging inside the cost, discount capped at 10% and only on dishes under 28% food cost.

If delivery passes 40% of sales, stop thinking like a restaurant with delivery and start running a separate unit, with its own kitchen and its own break-even; the global cloud kitchen market reached USD 80,300 million in 2025 according to Grand View Research, and that economy already plays by its own rules. Open your P&L this week and split the channel. Whatever number comes out decides everything else. The first difference is accounting, not marketing. The traditional method costs the Rappi dish with dining-room food cost and assumes the commission somehow comes out of profit; the Masterestaurant method places it INSIDE the item cost, alongside packaging and transit shrink, which on hot saucy dishes runs between 3% and 6%.

The four differences that decide your margin

A 30,000-peso dish at 30% food cost, 1,900 pesos of packaging and a 26% commission returns 5,300 pesos of contribution, not 21,000. Whoever skips that arithmetic before publishing is giving away product with a smile. Second comes digital menu size, where longer is strictly worse. Aaron Allen, founder of Aaron Allen & Associates, has argued publicly for years that oversized menus destroy margin by multiplying inventory, waste and line time without moving the ticket. Delivery doubles the penalty, since every extra minute of prep degrades delivery time and the aggregator's algorithm quietly demotes your visibility. Twelve to eighteen well-chosen items outsell sixty badly chosen ones, and you can verify that in your own zone ranking within a month. Third: which lever you pull to sell more. Discounting is the fastest lever and the most expensive one, because every discount point comes straight out of contribution while every average-ticket point flows almost entirely in.

The four differences that decide your margin — in practice

Bundling a 6,000-peso beverage at 18% food cost produces more profit than two orders sold at 30% off. The recommendation engine inside the Masterestaurant ecosystem builds those bundles from what your own customers order together, not from an industry average that describes nobody. And fourth, the one almost nobody watches: decision speed. The traditional method prices once or twice a year; the Masterestaurant method reviews three items every week on a dashboard that already carries contribution calculated. Fifty-two small decisions against two large ones. In a business where input costs move monthly and aggregator commission plans change underneath you, review cadence beats the brilliance of any single isolated decision.