How to increase restaurant sales on Rappi: the mistakes that actually cost money, and the right method

Acceptance time is the metric that moves the most cash

Accepting the order in under 60 seconds is the cheapest lever a restaurant has inside Rappi, because the clock starts when the order lands, not when somebody walks past the tablet. That number feeds straight into your position in the local results list, and position decides everything else: the 37% of adults who order delivery at least once a week, according to UpMenu (Food Delivery Statistics 2024), almost never scroll past the first screen. Do the arithmetic of an ordinary Tuesday: 45 orders accepted at an average of 3 minutes means 135 minutes of accumulated penalty that the platform reads as unreliability. Put the tablet where the expediter can see it, with the volume up, and measure for two weeks. This is a kitchen layout decision, not a marketing one. A listing without a photo or a description converts worse and therefore sinks your ranking, while commission is only charged on what actually sold.

Why does an incomplete catalog cost you more than the commission?

That is the accounting trap: the owner fights over commission points and gives away conversion. Run it with real numbers. A restaurant billing 40 million pesos a month in the channel at 25% commission hands over 10 million;

add 1,200 pesos of packaging across 1,300 orders and that is 1.56 million more, so the effective food cost of the same dish at the same price jumps from 30% to 40%. Negotiating two commission points gives you back 800,000 pesos. Completing 60 listings with your own photo and a two-line description moves the whole volume. Platforms reward completeness because it earns them commission: that is aligned incentives, not a favor. Price the delivery menu from contribution margin after commission, never from the dining room price. I got this wrong for years, treating the channel as an extension of the room with the same price list.

Channel pricing works backwards from what you were taught

A 38,000-peso dish with 11,400 pesos of raw material leaves 30% food cost in the room; the same dish in the app, with 25% commission and 1,200 pesos of packaging, leaves only 15,900 pesos of gross margin against the 26,600 the room delivers. The gap is 10,700 pesos PER PLATE. And the context does not help: ACODRES reported in 2025 a 9.8% rise in dish prices in Colombia since February, with 98,000 jobs at stake. Raise the channel price between 12% and 18%, and say so plainly in the listing if the customer worries you. When an item runs out and nobody switches it off in the app, the restaurant does not lose one order: it loses the order, the rating and its place in the list. It is the quietest leak in the channel, because it shows up in no report — what gets cancelled is not billed, and what is not billed is never analyzed.

Availability and hours: the sale nobody records losing

Suppose your signature dish is 22% of orders and sells out three nights a month without being switched off: between cancellations and refunds, that location gives up roughly 66 orders and drags its rating for weeks, long after the supplier restocked. The operating rule is simple and fits in one shift: whoever receives the delivery holds the panel password, and switching an item off takes eleven seconds. Write it into the cash close, not into a manual nobody opens. Channel benchmarks do not apply the same way in a 40-seat room as in a nine-location group, and mistranslating them causes half the bad decisions. In a SMALL restaurant, under 300 monthly orders in the app, the focus is single: acceptance time and a complete catalog, because you have no volume to negotiate with and position is your only asset. In a MEDIUM one, 800 to 1,500 orders, it starts making sense to split out the channel P&L, set your own prices and measure actual prep time against the promised one, a figure the platform already hands you for free.

How to read these numbers in YOUR operation?

In a GROUP with more than 4,000 monthly orders, the conversation changes in nature: there you do have leverage on commission, and each point on 160 million pesos is worth 1.6 million a month.

The scale figures in this document come from public operator reports and market data houses, not from primary research. DoorDash marketplace volume — roughly US$80.2 billion in 2024, per its fourth-quarter results — and Just Eat Takeaway.com's GTV of EUR 26.3 billion in 2024 help size the sector, not predict your till. Statista Market Insights puts the worldwide grocery delivery segment at US$786.8 billion in 2024, a different order of magnitude altogether. Neither Rappi nor iFood publishes its ranking formula, so the five metrics prioritized here come from what both expose in the merchant panel, not from a leaked document. The peso amounts are illustrative and you must recalculate them against your own menu before touching a price.

The mistake of reading the channel without its own P&L

A restaurant that does not split out the delivery P&L discovers the problem at year-end, and by then it has spent twelve months subsidizing orders. Diego F. Parra insists on this in the Masterestaurant method because the arithmetic is merciless: commission, packaging, waste from remakes and the kitchen time stolen from the dining room never appear on the sales line, only on the profit line. Build a sheet with four columns — channel sales, commission, packaging, raw material cost — and calculate contribution margin per dish, not per business. You will find two or three items losing money on every order, almost always the large-portion ones with the room price untouched. Pull them off the app or reprice them. That cleanup alone usually beats a full month of in-app advertising. If Rappi dropped you two positions in your zone tomorrow and you lost 40% of channel orders, what would be left?

What if the app switched off 40% of your orders tomorrow?

That question has an accounting answer and it pays to calculate it before you need it. A location with 1,300 monthly orders and a 42,000-peso ticket bills 54.6 million in the app;

losing 40% is 21.8 million in sales and, at a 42% contribution margin after commission, some 9.1 million pesos of monthly profit that vanish without warning. That is the paradox of the channel: it gives you volume while taking your customer, because the contact data is not yours. The way out is not abandoning the app — that is ideology, not management — but using it for acquisition and moving repeat buyers to your own delivery with an insert in the packaging. Start this week: measure your acceptance time over the last 30 days. The first difference is accounting, not marketing. A restaurant selling 40 million pesos a month on Rappi at 25% commission hands 10 million to the platform; add 1,200 pesos of packaging across 1,300 orders and that is 1.56 million more.

The five differences behind 90% of the gap

This channel needs its own P&L, because the food cost you calculated for the dining room no longer applies: the same dish at the same price jumped from 30% to 40% effective food cost. Whoever skips the split never sees the problem until the year closes badly. Second comes the catalog. Delivery platforms reward completeness —photo, description, category, availability— because a complete catalog converts better and earns them more commission. That is not a favor; it is aligned incentives. A restaurant with 100% original photos and short descriptions typically sees two to three times the conversion of a catalog with half the items blank, and that jump costs no ad budget, just one afternoon of shooting in window light. Third is operational and the most thankless: timing. Accept fast, cook within the time you promised, do not cancel. Rappi and iFood measure all three and turn them into position.

The five differences behind 90% of the gap — in practice

A restaurant cancelling 6% of its orders because the catalog availability was never synced pays for that carelessness in visibility, and lost visibility does not come back with a coupon. Fourth is portfolio. A virtual brand —a concept that exists only inside the apps, cooked in the kitchen you already pay for— multiplies your search surface without new square meters. If your pasta restaurant fights 90 Italian competitors in its zone, a healthy-bowls brand drops you into a category with 20. The dark kitchen vs physical restaurant debate usually resolves right there: squeeze the kitchen you have first, rent a new one later. And the fifth one is judgment, the kind no dashboard shows: deciding what NOT to upload. The most expensive mistake I keep running into is uploading the dining-room signature dish —the one served hot, crisp, plated to order— and letting it arrive lukewarm 25 minutes later. That dish destroys the rating, and the rating destroys position. Decide first which dishes do NOT travel; a delivery menu is built by subtraction, never by copy.