Selling on Rappi in Colombia: what the restaurant must know, in 2026 numbers

What do you actually keep from a 40,000 COP Rappi order?

From a 40,000 COP Rappi order on a full-logistics plan you keep less than 4,000 COP in contribution, and the kitchen that cooked it still has to come out of that.

The arithmetic allows no argument: a 25% commission takes 10,000 COP, hot-food packaging in Colombia runs between 1,800 and 3,200 COP depending on format, and a dining-room food cost of 33% swallows another 13,200 COP. Add it up and 3,600 COP remain to pay the cook, the gas, the electricity and the rent on those square meters. Colombia's digital channel already moves close to 22% of urban restaurant sales according to the Asociación Colombiana de la Industria Gastronómica, so we are talking about a fifth of the till operating on subsistence margins. The decision these numbers force is simple and harsh: either drop the delivery menu's food cost below 28%, or push the average ticket above 45,000 COP, or the channel is financing volume with your own profit.

Commission is not a number, it is a plan you choose

Rappi's commission in Colombia moves by plan rather than a single rate, and that is exactly where most owners lose money without noticing. Basic schemes favor the restaurant that absorbs part of the logistics, while plans with promotional visibility and full logistics cost more. Twelve points of difference on monthly sales of 24 million COP means 2.88 million COP every month, more than the rent of many neighborhood locations in Bucaramanga, Pereira or Manizales. What almost nobody calculates is the return on those twelve points: if the expensive plan brings 40% more orders, you win; if it brings 12%, you are buying visibility you already had. Measure three months on the basic plan, three on the premium one, and compare CONTRIBUTION in pesos, not order counts. This is a portfolio decision, not a negotiation: pick the plan that leaves the most cash after commission, not the one showing the most orders on the dashboard.

The tax mistake that blows up in your first two months

VAT and withholding tax are calculated on the gross price the diner paid, not on what the aggregator deposits into your account, and that accounting detail has wrecked the cash projection of more than one Colombian restaurant in its first fiscal bimester. If the customer paid 50,000 COP and Rappi settled 36,000 COP net, the DIAN will ask you about the 50,000 COP. An owner budgeting from the settlement net is underestimating the tax burden by a double-digit percentage, and finds out once the money is spent. This is where the MASTERESTAURANT method separates two ledgers: the aggregator settlement reconciles treasury, and only the gross sales report drives tax provisioning. Add a fixed line to your weekly cash flow called digital channel tax provision, feed it from the gross figure, and never touch it. Without that line, your channel profit is an accounting mirage that lasts exactly sixty days.

The algorithm rewards speed, not excellence

Rappi, Uber Eats and DiDi Food rank restaurants by conversion, preparation time, cancellation rate and rating, in that order of operational weight, which is why a restaurant with 4.3 stars and 14 minutes of prep shows above one with 4.8 stars and 31 minutes. The aggregator does not optimize diner happiness: it optimizes completed deliveries per courier hour, because that is its dominant variable cost. There is a paradox of the trade worth resolving head-on here. Your best dish, the one holding up the dining room's reputation, may be precisely the one sinking your ranking if it takes 28 minutes on the grill. The fix is not lowering quality: it is pulling that dish from the channel and keeping it as a table exclusive. Cut the digital menu down to whatever leaves the pass in under 15 minutes with a full kitchen on a Friday at 8 p.m. That pruning buys more visibility than any promotional plan.

Redesign the delivery menu: its own costing, its own price

A delivery menu carrying dining-room prices is a costing error, not a commercial strategy problem, and it is the root cause of that 22% of digital sales leaving no profit. The engine's rule is plain: maximum food cost per dish is 32% and even that is not recommended, so in a channel that charges commission you need to bring that food cost down to breathe. Three routes get you there and none of them is raising prices blindly: reformulate portions for the actual container, drop garnishes that travel badly and nobody pays for, and lift the digital price by 10% to 15% while showing it openly. The Colombian diner already understands that delivery costs more; what gets punished is the surprise, not the premium. Close the exercise with a separate costing sheet where packaging is a direct cost of the dish and not an administrative expense lost somewhere in the month.

Packaging: the line item that does not exist in the dining room

Packaging a hot-food order in Colombia costs between 1,800 and 3,200 COP, a line item that simply does not exist when you serve on porcelain, and on a 35,000 COP ticket it represents between 5.1% and 9.1% of the sale. Set it beside the commission and the full problem appears: a 25% plan plus 7% packaging equals 32 points of the sale before touching the first ingredient. Diego F. Parra puts it bluntly during Masterestaurant audits: packaging gets costed per dish, with its reference, its supplier and its unit price for the month, exactly like the protein. Switching a container from 2,900 COP to 2,100 COP across 900 monthly orders recovers 720,000 COP in profit without touching the recipe or the price. The decision these figures produce belongs to purchasing, not to the kitchen: negotiate packaging on quarterly volume and verify the unit cost every single month.

Virtual brand or your own delivery? The numbers at the crossing

A virtual brand inside the aggregator makes sense when your kitchen has idle capacity in identified time slots, and the cloud kitchen market is growing at 12.6% annually between 2026 and 2033 according to Grand View Research, while the global dark kitchen segment advances at 12.7% annually between 2025 and 2033 per Global Growth Insights. That expansion speed means fresh competition every quarter inside your own coverage zone. Your own delivery fleet, by contrast, starts paying once the digital channel passes roughly 400 recurring monthly orders from customers who already know you: at that point a courier's cost spreads across enough orders to fall below the 18% the cheapest plan would charge. Below that volume, building your own logistics swaps a variable commission for a fixed cost you cannot switch off on a rainy Tuesday in February. Measure your recurring orders before buying the first motorcycle.

The 3 numbers you should tattoo on yourself

28% food cost on the digital menu, a 45,000 COP average ticket and 15 minutes of preparation: those three figures decide whether Rappi is a channel or a leak. For the first, the action is a separate delivery costing sheet this very week, where packaging at 1,800 to 3,200 COP enters as a direct cost of the dish and every recipe above 28% either leaves the channel or gets reformulated. For the second, build combos and suggested add-ons that push the ticket above 45,000 COP, because the same 25% commission hurts far less on a high ticket than on one of 30,000 COP. For the third, time your line with a full kitchen on a Friday night and delete from the digital menu everything above 15 minutes, so the algorithm lifts you up the list. Start today with the first one: without your own delivery costing sheet, the other two are guesswork.

The numbers that change the decision

Rappi commission in Colombia moves by plan: basic tiers favor restaurants that carry part of the logistics, while full visibility and courier packages cost more. On monthly sales of 24 million COP, those twelve points equal 2.88 million COP, more than the rent of many neighborhood venues in mid-sized cities. VAT and withholding are calculated on gross, not on what the aggregator wires you. An owner who budgets against the net figure on the settlement statement gets a nasty surprise in the first tax cycle, because the authority will ask about the price the diner actually paid. Packaging for a hot Colombian order runs 1,800 to 3,200 COP depending on format.

The numbers that change the decision — in practice

That line item does not exist in the dining room and almost nobody folds it into channel costing, so it shows up as a blurry leak in the P&L filed under supplies. Platform promotions are co-funded. A 2x1 featured on the app's home screen can cost you 40% to 60% of the discount value, and that spend hits your margin this month, not the platform's marketing budget. Prep time outweighs rating in the ranking. Cutting the promise from 28 to 16 minutes usually moves more orders than climbing from 4.4 to 4.7 stars, because the system allocates by completed deliveries per courier hour. A well-built virtual brand inside the same kitchen can add 15% to 35% of incremental sales without a peso more of rent, provided it uses ingredients you already buy and adds no station to the hot line.