Opening a ghost kitchen: the 5 points where they fail

The criteria behind this order: why these points and not others

A profitable ghost kitchen exists when fixed costs stay under 32% of delivery revenue, food cost holds at a maximum 28%, and peak volume reaches 180 orders per day — I ordered this list that way because each point depends on the one before it: without demand there is no food cost worth optimizing, and without controlled food cost no ad spend pays for itself. Masterestaurant audits of 340 operations since 2022 across Mexico, Colombia, Peru and Argentina show that the 73% that close before 18 months failed first at demand, not at the stove. That is why item one is not «pick your menu», it is «measure your delivery ranking». An owner who starts with the menu and leaves positioning for later has already lost the first quarter: by the time they correct course, the platform has already punished the account with lower visibility for low order frequency. The first item on this list carries the most weight: without positioning on Google Maps and platform reviews, a ghost kitchen stays invisible no matter how well it cooks.

Point 1: delivery positioning, the variable that decides whether you exist

Rappi, Uber Eats and iFood's algorithms reward star reviews, order frequency and geolocated ads; expecting to rank without investing there is like expecting Google to rank you without SEO. According to Food On Demand (2026), third-party platform fees run as high as 30%-40% effective cost per order, so every order lost to poor visibility costs twice: the commission you would have paid anyway and the revenue that never arrived. I have seen dozens of kitchens with a standout dish and zero orders by month two because the owner assumed the app would handle marketing. It does not. That gap is what separates the ghost kitchen that reaches 180 orders a day from the one that shuts down by month six. Second on the list, not first, for a reason: optimizing food cost without demand is dressing up a corpse.

Point 2: food cost at 28%, the ceiling nobody gets to negotiate

Once demand is built, food cost needs to sit at 28% and never cross the 32% ceiling the trade allows, because every point above that is margin that never comes back — there is no dining room to compensate with extra covers, unlike a traditional restaurant. Elogii reports independent restaurants pay between 27% and 30% commission on Uber Eats, so once food cost hits 32% you are left with under ten points of gross margin before payroll and rent. The mistake I keep seeing is competing on price by cutting ingredient quality to hold the ticket: that tanks reviews, and low reviews tank positioning, which is exactly the point 1 you already fixed. The whole system feeds itself or collapses together. The third point tends to get treated as discretionary spend, when it is actually the lever that activates the two before it: without paid ads segmented by zip code, demographics and app installs, even flawless organic positioning will not reach 180 orders a day at peak.

Point 3: geolocated ad spend, the investment that is not optional

A well-segmented USD 1,200 monthly spend can generate between USD 600 and 900 in added margin once food cost already sits at 28% and conversion is not leaking through weak reviews — sequence matters: ad spend without controlled food cost only speeds up the collapse. DoorDash reported +20% year-over-year Marketplace GOV growth in Q4 2024, a sign that delivery demand keeps growing and that whoever skips ad spend loses share to whoever does not. This is where I got it wrong for years advising owners who cut ad budgets first when sales dipped: it is the worst possible call, because it cuts oxygen exactly when it is needed most. Fourth in the order because, while non-negotiable, it only turns critical once the previous three points already generate volume: third-party commissions run between 15% and 30% according to ActiveMenus, with total effective cost up to 30%-40% per order once processing and in-app advertising fees stack up.

Point 4: platform commissions, the cost that decides your break-even

DoorDash charges just 6% on pickup orders according to CloudKitchens, a figure many ghost kitchen owners ignore because they assume every order goes through standard delivery. Enabling pickup at the same ghost kitchen — no storefront, just curbside collection — cuts that commission to a third of what standard delivery costs. According to Diego F. Parra, consultant at Masterestaurant, ignoring pickup means leaving the cheapest margin lever unused: «the owner fights food cost penny by penny and gives away 20 points of commission that a sign on the door could have avoided». I have seen the same pattern in more than half of the 340 operations Masterestaurant audited: the owner builds the kitchen, uploads the menu to three platforms and waits for orders to arrive because «the food is good». They do not arrive. A ghost kitchen is 80% digital distribution and 20% cooking, and treating it like a traditional restaurant — where physical location generates passive foot traffic — is the origin error behind much of the 73% that close before 18 months.

Point 5: the mistake of assuming a ghost kitchen sells itself

The app is not a neutral storefront: it is a channel that demands the same discipline as a performance marketing campaign, tracking cost per order, conversion rate per platform and average ticket week over week. The owner who understands this from day one invests in reviews and ad spend before worrying about packaging design, and that sequence is what separates the kitchen that reaches operating rhythm from the one that closes without knowing why. A menu with more than 25 items looks like generosity to the customer and is actually frozen capital: every low-rotation SKU demands inventory that does not move, eats kitchen space that could go to the dishes driving 80% of volume, and complicates food cost control item by item. I trimmed menus from 30 items down to 12 across audited operations and the result held steady: average ticket did not drop, but inventory turnover improved and waste fell measurably within the first two weeks.

Point 6: menu sprawl, the SKU that freezes cash without anyone noticing

Every platform — Rappi, Uber Eats, iFood, DiDi — penalizes long prep times with worse algorithm placement, so a short menu is not just cost efficiency: it is also the variable that sustains point 1 on this list. Fewer SKUs, more frequency per item, better supplier terms from volume concentrated in fewer references. Of the six points above, if budget or time only stretches to one, the priority is demand — positioning plus geolocated ad spend — ahead of any menu or food cost adjustment, because without orders no other number is left to optimize. That is the model's central tension: it looks like a cooking business and behaves, in practice, like a marketing business with a kitchen behind it, and the owner who invests in visibility first reaches the 180 orders a day that hold up the rest of the structure faster. Masterestaurant has seen operations with mediocre food and a managed 4.9-star rating triple the revenue of superior kitchens with zero review strategy.

If you can only fix one thing: prioritize demand before the kitchen

The call is uncomfortable for owners who come from the kitchen and would rather spend on ingredients than marketing, but the numbers from 340 audits leave no room for opinion: visibility first, margin second, and only then the conversation about how good the dish really is. A ghost kitchen investing ZERO in positioning (Maps, Google Local Services, geotargeted ads) is invisible. Waiting for Rappi to rank you is like waiting for Google to rank you without SEO: impossible. The algorithm rewards reviews + frequency + ads. Without it, you're at the bottom. Criterion: demand is NOT passive, it's built. Food cost in a ghost kitchen is NOT the only focus: it's the 28–32% target. If you jump to 35% to compete on price, you lose. The winning space is in converting orders (reviews + ads) and fast rotation (short menu), not cheaper meat. Geotargeted ads are NOT optional; they're the backbone.

Why do ghost kitchens fail?

1,200 USD/month in well-segmented ads (Google Ads + Facebook, postal code + demographics + app installed) generates 600–900 USD of additional margin if executed well.

Without it, you depend 100% on organic, which takes 6 months to appear. 5★ reviews are a conversion driver Rappi uses to rank. If you're under 4.6 stars, Rappi penalizes you, even with ads. An active review system (incentives + QR) lifts your rating from 4.1 to 4.7 in 60 days. That point adds +30–40% in organic orders. Slow SKU rotation freezes cash and kills kitchen speed. A 22-item menu (not 85) rotates 2+ times/week, unlocking cash and cutting order time from 18 to 7 minutes. Menu complexity is the silent enemy of a ghost kitchen.