How to increase restaurant sales on Rappi: before vs after with Masterestaurant

Step 1 · Build the channel P&L before you touch a single photo

Your first deliverable is not a promotion, it is a spreadsheet with the real margin per dish inside Rappi, and without it everything else is guesswork. Take the twenty dishes that move most on the app and subtract from each price the channel commission —between 18% and 30% depending on country, category and whether you pay for visibility—, the packaging, the average discount you ran last month and the dish food cost. What remains is your digital contribution margin. A 12 dollar dish that leaves 6.20 in the dining room often ends at 1.80 once it comes through an aggregator, and two or three usually show up negative. It is done when you can point at the table and name your five profitable dishes and the ones you are giving away. Verify it by matching the Rappi Partners biweekly settlement against your own sales report: both figures must reconcile to the dollar.

Step 2 · Cut the menu to 24 dishes that survive 25 minutes on a motorbike

The channel menu is not the dining room menu, and that cut is the cheapest conversion lever you own. A table dish is designed to arrive in ninety seconds; a delivery dish travels twenty-five minutes inside a thermal bag on a motorbike, with vibration and condensation. Risotto does not travel. Fries without double cooking arrive limp and cost you a three-star review that takes forty good orders to offset, and that is not a metaphor: Harvard Business School measured that each additional star in the rating moves between 5% and 9% of revenue. Pick twenty-four items, kill the other sixty-two, and test each one packed, waiting half an hour before opening it. The deliverable is a signed list of live and dead dishes. Verify it with the bag test: if you would not eat it, neither will your customer. This is where nearly everyone stops, and where half the profitability problem sits.

Step 3 · Raise the channel price 15% to 20% without guilt

A channel price 15% to 20% above the dining room deceives nobody: it acknowledges that the order carries an 18% to 30% commission, 0.60 dollars of packaging and zero table turnover, and that the customer is paying not to leave home. Rappi allows its own price lists per channel, so the work is operational, not strategic. Large United States chains already raised menu prices 42% between 2020 and 2025, almost double the 22% general inflation according to One Haus, and they are still full. The owner's resistance is psychological, not market driven. It is done when the Rappi Partners panel shows the differentiated list active. Verify it by measuring conversion two weeks later: if it drops less than three points, you left money on the table. The prep time you declare in the app is a contractual promise, and breaking it costs you ranking and reviews at the same time.

Step 4 · Bring prep time to 12 minutes and publish what you actually meet

Time thirty real orders with a stopwatch, from the tablet chime until the courier walks out with the bag sealed, and keep the ninetieth percentile, not the average. If that number is eighteen minutes, declare eighteen, not twelve: the algorithm punishes the miss harder than it rewards the optimistic promise. To truly bring it down you need a prep line separate from the dining room ticket, pre-assembled packaging at station and one cook dedicated to the channel during the two peak hours. In the digital channel audits we run at Masterestaurant, the bottleneck almost never sits in the kitchen but in assembly and bag sealing. The deliverable is a declared time you meet on nine out of ten orders. Your customer decides inside the app in under forty seconds, and that window is won by the photo or lost by the copy. The images your beverage supplier shot in 2019 are killing conversion on your most profitable dishes.

Step 5 · Photography and listing: forty seconds to decide

Shoot the eight highest-margin dishes —the ones that came out profitable in Step 1— with natural light, a clean background, the plate filling 80% of the frame and the actual packaging the customer will receive, not porcelain they will never see. The description stays under two hundred characters, with weight and sides stated explicitly, because half of all refund claims are born from an expectation you failed to clarify. The deliverable is eight rebuilt listings. Verify it by comparing the conversion rate of those eight against the prior month in the panel: if it does not climb at least four points, the photo is still your problem. The most expensive mistake is not pricing, it is accepting every promotion your account executive offers without running it through the Step 1 margin. A two-for-one on a dish that leaves 1.80 dollars puts you to work for free and paying for the privilege.

The four mistakes that will cost you the quarter

The second mistake is buying visibility before fixing the photo and the time: pushing traffic toward a bad listing is burning budget with style. The third, and I see it over and over, is switching the channel off at peak hour to avoid missing times, which the algorithm reads as low availability and which sinks your position for weeks. The fourth is copying the dining room price out of fear, and that is exactly where the phrase «I sell more and I have less» comes from. Each one has the same antidote: numbers before decisions, always in that order. Run the scenario all the way out and the decision makes itself. If you keep the full menu, the dining room price and the optimistic time, gross channel sales will grow because demand is there —Statista projects 1.89 trillion dollars in food delivery for 2026, and Rappi reports more than 45 million active users in Latin America in its public 2025 figures—, but your contribution margin dilutes quarter after quarter.

What happens if you change nothing?

The kitchen works longer hours, the staff wears down running two services at once, packaging eats the differential, and by year end you have 30% more billing and less cash than the year before.

That is when the owner blames the commission. The commission is a known, published cost from day one: the real problem was switching on a channel without giving it a price of its own. You are finished when you can answer five questions with a number rather than a feeling. First: what is the average contribution margin of your digital menu after commission, packaging and discounts? It should sit above 30%. Second: how many dishes are published? Twenty-four or fewer, with food cost under 28% on the ones that move most. Third: is your channel price between 15% and 20% above the dining room across the whole menu, with no forgotten exceptions? Fourth: what share of orders meets your declared time?

Closing checklist: how you know it came out right

Nine out of ten, or go back to Step 4. Fifth: did your rating rise or fall over the last thirty days? Open tomorrow's Rappi settlement for the latest fortnight and calculate the real margin on your three best-selling dishes. That figure, not gross sales, tells you whether the channel is working for you. THE MENU is not the same menu. Dine-in is engineered for a 90-second walk to the table; the channel has to survive 25 minutes inside a thermal bag on a motorbike. Risotto does not travel. Fries without a double fry arrive limp and cost a three-star review that takes 40 good orders to offset. Keeping 24 dishes and killing 62 sounds like amputation and works the opposite way: a short menu lifts conversion because the user decides in under 40 seconds inside the app. PRICE has to differ, and that is exactly where the owner gets nervous.

The four differences that move the number

Channel pricing 15-20% higher deceives nobody: it acknowledges that the order carries an 18-30% commission, $0.60 of packaging and zero table turns. Rappi supports channel-specific price lists; an operator who ignores them is subsidizing the aggregator out of their own margin. The honest objection —«what if the customer compares and gets angry»— has a measured answer: delivery elasticity is low, because the user benchmarks against cooking at home, not against your dining room. TIME is the heaviest variable in the algorithm and the least touched. Rappi ranks by likelihood of conversion, and that depends on distance, rating and estimated time. Dropping prep from 24 to 13 minutes pushes the restaurant up its category without a single peso of advertising. It is achieved through boring work: separate mise en place for the channel, an assembly station with packaging within reach, and a ban on the channel cook covering the dine-in pass during peak.

The four differences that move the number — in practice

THE PRODUCT LISTING converts. Owned photo, portion weight, a two-line description naming the differentiating ingredient, declared allergens. Sounds cosmetic until you look at view-to-order conversion: items with an owned photo and a complete description convert around twice as well as name-and-price listings. And a concession here: for years I filed this under soft marketing and I was wrong; it is pure operations, because a faithful photo cuts the «this is not what I ordered» complaints that wreck ratings.