Content-to-Table Conversion Index 2026: selling on Rappi or iFood from a dark kitchen, when it is worth it

Finding 1 — What does selling an order on Rappi or iFood from a dark kitchen actually cost?

It costs between 35% and 45% of the order value once every surcharge is added, not the 15%-30% printed in the contract, according to CloudKitchens (2026).

Rezku (2026) documents that the NOMINAL commission charged by DoorDash, Uber Eats and Grubhub sits in that 15% to 30% band per order, and that is the figure a ghost kitchen owner writes into the cost sheet when deciding to join the channel. Trouble shows up afterwards, with payment processing surcharges, co-funded promotions, cancelled-order adjustments and the price of ranking high on the list, which CloudKitchens (2026) adds up until the EFFECTIVE commission reaches 35%-45%. A dish carrying 30% food cost and 40% effective commission leaves the operator the remaining 30% to cover packaging, kitchen payroll, shared-space rent and utilities. That margin does not fund an operation; it funds the illusion of revenue. A dark kitchen owner in Medellín billed 92 million pesos a month on Rappi and still could not pay himself a wage, because his problem was never sales, it was OWNERSHIP of the channel.

Finding 2 — The Medellín case: 92 million pesos a month and no salary for the owner

Every peso came in through a door that was not his, and went out through a commission he neither set nor could negotiate. Apply the band CloudKitchens (2026) reports to that revenue and somewhere between 32 and 41 million pesos evaporated before touching a single ingredient. Diego F. Parra and Masterestaurant have spent years repeating the same reading to operators who swear they have a food cost problem: their dishes cost what they should, prime cost is under control, and yet break-even drifts further away every month because the channel mix keeps getting worse with nobody keeping score. The underlying error consists of treating commission as a marketing expense instead of what it is, a share in the ownership of your customer. The app earns its place when it works as a paid DISCOVERY channel rather than a repeat-order channel, because Lightspeed (2025) measures that 42% of diners use third-party apps only to reorder from a restaurant they already knew.

Finding 3 — When the app does pay off: the three scenarios that survive the arithmetic

That 42% describes traffic you already earned and are now paying for twice. The arithmetic holds in three concrete situations. First, a brand-new kitchen with no name and no database, where the 35%-45% reported by CloudKitchens (2026) buys the first three hundred orders and with them the first three hundred phone numbers. Second, off-peak hours with idle installed capacity, where the marginal cost of one more order is just the input. Third, proving out a virtual brand before committing capital. Outside those three cases, every app order that repeats month after month is perpetual rent charged on a customer who is already yours. The Content→Table Conversion Index measures the path from someone seeing local content to placing a DIRECT order, and that metric decides whether you pay 35%-45% effective commission (CloudKitchens, 2026) or the variable cost of a geotargeted ad. This is no vanity number, it is the difference between renting demand and building it.

Finding 4 — The metric kitchens without a dining room write off

BrightLocal (Local SEO Statistics 2026) reports that 64% of US diners search Google for restaurants before visiting, and The Media Captain (2024) adds that 42% of local searches end in a click on the local pack, that map block with three listings. A dark kitchen with no visible address forfeits that 42% by default, unless it works its business profile with a declared delivery zone. What you recover along that path is not merely the order margin, it is the diner's data, their history and their frequency. TouchBistro Diner Trends (2025) measures that 25% of diners would avoid a restaurant because of criticism on social media, a share no amount of paid placement inside Rappi will offset. The same study establishes that 41% research on social platforms where to eat before deciding, and Restroworks (2025) raises the figure to 62% when it comes to checking the restaurant's page before choosing.

Finding 5 — Reviews weigh more than your ranking inside the app

Taken together, the public conversation about your brand decides more orders than the app's algorithm does, and you control the first while renting the second. BrightLocal (Local Consumer Review Survey 2025) supplies the operable number: 96% of consumers are willing to write a review. Almost nobody asks. One message the day after the order, naming the dish and carrying a direct link, turns that 96% willingness into reputational inventory of your own, which no app can charge you for or take away when you decide to walk. The operating goal consists of migrating the diner from the app to your direct channel before the third order, and the measured levers are SMS, email and the loyalty point. Sakari (2025) reports that 84% of consumers opted in to receive SMS from at least one business, and Tabular (2025) measures that abandoned-cart SMS messages in restaurants hit between 10.1% and 14.2% click rates, numbers no in-app banner matches.

Finding 6 — How you win back the customer the app sent you

Stripo (2025) puts at 55% the diners influenced by quality promotional emails. And the National Restaurant Association (2025) closes the argument with two figures: 78% of consumers are more likely to visit if they earn points, and nearly 90% would use exclusive offers from a restaurant's own app. Against those numbers, the 15% discount you give away on your direct channel still runs half as expensive as the 35%-45% effective commission CloudKitchens (2026) reports. Here sits the tension nobody resolves out loud: the more volume Rappi or iFood brings you, the harder walking away becomes, because your break-even was already calibrated against revenue inflated by commission. Picture a kitchen with 60% of income coming from apps at a 40% effective commission per CloudKitchens (2026). Cut the channel tomorrow and you lose that 60% at once, while your shared-space rent drops by nothing. Keep it and your margin never grows.

Finding 7 — The channel paradox: whoever sells most through the app can least afford to leave it

The exit is neither binary nor instant, it gets built by shifting three points of share per month toward the direct channel while the app carries the volume. Lightspeed (2025) supplies the underlying argument: customers who order online visit 67% more frequently. That frequency is the asset, and today a third party is capitalizing it and charging you rent for the loan. Calculate your real effective commission by dividing what the app deposited last month by the gross value of those orders, and if the result clears the 35% CloudKitchens (2026) marks, you have a structural decision pending, not a menu tweak. Against that number, set the cost of acquiring a direct order, which in most kitchens I review lands between 8% and 15% of the ticket once you count local ads, SMS and the loyalty point. Morning Consult and NRN (2025) also remind us that 37% of Americans dine out less frequently, so the market will not grow to cover the hole the channel leaves.

Finding 8 — The number that decides, and what to do on Monday

Open the sheet on Monday, run that division and write the percentage in large type. If it comes back 41%, you already know how much of your kitchen belongs to somebody else. SOURCES SYNTHESIZED. This analysis contrasts six public sources: Rezku (2026) and CloudKitchens (2026) for the real cost of the third-party channel; Lightspeed (2025) for online ordering behavior; BrightLocal (Local SEO Statistics 2026 and Local Consumer Review Survey 2025) for discovery and reviews; The Media Captain (2024) for how local clicks split; the National Restaurant Association (2025) for loyalty and app offers; plus Tabular (2025), Stripo (2025), TouchBistro Diner Trends (2025) and Morning Consult/NRN (2025) for recovery levers and frequency context. SELECTION CRITERIA. Only sources publishing methodology or sample size made the cut, dated between 2024 and 2026, each reporting a figure an owner can act on: a commission percentage, a conversion percentage, a declared behavior percentage.

Finding 9 — Sources, scope and method of this synthesis

Vendor reports that publish a headline without any breakdown were dropped. Where two sources disagree — headline commission against effective commission — BOTH are cited and the gap is explained, rather than averaged into one comfortable number. TIME WINDOW. Data spans 2024 to 2026, weighted toward 2025. That window matters because the diner frequency adjustment — the 37% dining out less often per Morning Consult/NRN (2025) — reorders the entire channel calculation and makes any pre-2024 benchmark irrelevant. WHAT MASTERESTAURANT ADDS. The numbers belong to the cited sources. What Diego F. Parra and Masterestaurant contribute is the READING: organizing those figures into a single measurable content-to-table path, defining the channel dependency threshold, and setting the order in which an owner should attack the problem without breaking break-even mid-migration. HONEST LIMITATIONS. First, most of these sources measure the United States market, while Rappi and iFood commissions across Latin America vary by country, category and commercial agreement, so the cited ranges orient you but never replace your own contract.

Finding 10 — Sources, scope and method of this synthesis — in practice

Second, several figures are DECLARED survey behavior rather than observed at the till, and declared intent always flatters itself. Third, no serious comparable public source isolates pure dark kitchen performance against kitchens with a dining room, so that cut is a reading on our part, flagged as such, and we do not invent a datapoint nobody published. OPERATIONAL DEFINITIONS. Content-to-Table Conversion Index: direct orders or bookings attributable to local content divided by local impressions in the period, as a percentage. Effective commission: total retained by the app (commission plus surcharges plus in-app advertising plus co-funded promotions) over gross order value, as a percentage. Channel contribution margin: average ticket minus food cost minus variable channel cost, in currency per order. Channel dependency: third-party sales over total sales, as a percentage. Diner LTV: contribution margin per order multiplied by annual frequency and by customer lifespan in years.