Delivery Unit Economics: Myth vs Reality in Restaurants (2026)

The myth: delivery is "extra money" that drops straight into the till because the dish was already costed for the dining room. The reality: once you add the app commission (25%-30%), packaging (3%-5% of the ticket), the dish's food cost (which must stay at or below 32% to protect margin) and the kitchen time lost on low-ticket orders, the real unit economics of a delivery order lands between -4% and +6% net margin, depending on channel mix. At Masterestaurant we've audited more than 180 restaurants across Latin America, and the number that shocks owners most is this: 6 out of 10 dishes sold through delivery apps run at a loss once you load the real packaging and commission cost onto them. The mistake isn't selling through delivery; it's never re-costing the menu for that channel. Here's the full math, with verifiable figures, so you decide with numbers instead of gut feeling.
The myth was born in 2020-2021, when delivery jumped from 8% to 35% of sales at thousands of restaurants forced to shut their dining rooms. Back then, any app sale felt like incremental revenue on top of an operation already paying rent, payroll and fixed utilities. The problem is that logic froze in time: commissions from delivery apps climbed noticeably from the rate charged when the channel first launched across most Spanish-speaking markets.
Diego F. Parra puts it bluntly: 'delivery stopped being the pandemic's extra income; today it's a channel with its own P&L, and if you don't cost it separately, it's draining margin you'll never see on the general income statement.' Treating delivery as a dining-room sale with a box on top is exactly where the invisible loss starts, and Masterestaurant sees this pattern in nearly every audit we run across full-service and dark kitchen operations alike.
Side-by-side: delivery unit economics
| Myth (what the owner believes) | Reality (2026 unit economics) | |
|---|---|---|
| App commission | ✕Commission from the channel that cuts into the margin versus dine-in sales. | ✓25%-30% real, up to 35% with sponsored ads |
| Dish food cost | ✕28%, same as dine-in | ✓34%-37% real without re-costing packaging and portion |
| Packaging and bags | ✕1% of the ticket, a minor cost | ✓3.5%-5% of the ticket in operations with 200+ orders/month |
| Kitchen time per order | ✕8 minutes, same as dine-in | ✓11-14 minutes real, due to friction and double-checks |
| Net margin per order | ✕+15% net, same as dine-in | ✓-4% to +6% real depending on channel and ticket |
| Average ticket | ✕$180, same in both channels | ✓Lower ticket than dine-in, pressured by the channel's commission. |
Delivery is not incremental revenue: it has its own P&L
The delivery channel generates a net margin of 8% to 14% when costed independently; if you fold it into the dining room P&L, that number vanishes between shared line items and looks like profit when it is actually eroding margin. The myth was born in 2020-2021, when delivery jumped from 8% to 35% of sales at restaurants forced to close their dining rooms: every app order felt like money on top of an operation whose rent and payroll were already covered. That reasoning got frozen in time. By 2026, commissions from delivery apps at Latin American urban restaurants climbed noticeably above the rate charged when the channel launched. Diego F. Parra repeats it in every audit: 'If you don't have a separate P&L by channel, you have no idea whether delivery is giving you money or taking it.'
Base commission plus sponsored ads: the real cost exceeds 30%
The base fee charged by delivery apps in Spanish-speaking markets ranges from 25% to 30% of the ticket, but climbs another 3%-8% when the restaurant activates sponsored ads inside the platform to capture top-of-feed visibility. A restaurant with a $18 USD average ticket pays between $4.50 and $6.84 in commission plus advertising before touching the cost of a single ingredient. In practice, operators who keep sponsored placements running permanently — common among dark kitchens competing in saturated feeds — can reach an effective commission of 35%-38%. Masterestaurant recommends always calculating the «total effective commission» before setting app menu prices: add the base fee, the ad spend percentage and any zone-adjustment charges, then apply food cost against the resulting price — not against the dine-in price.
Delivery food cost: the same dish can run 6-9 points higher
A dish at 28% food cost in the dining room can reach 34%-37% in delivery if the portion size, extra sauces and packaging are not re-engineered for the channel. The reason is structural: on the line, the cook portions by eye under the pressure of service and the chef corrects in real time; in delivery, the portion goes out without that visual check, and sauces tend to double up 'so the customer doesn't complain,' which drives noticeable portion creep on dishes that lack a dedicated delivery-adapted standard recipe. Applied to a raw material cost of $4.20 USD per dish, that extra 10% is $0.42 per order; multiplied by 180 monthly orders, it adds up to $75.60 USD in silent monthly leakage — from a single menu item alone.
Packaging consumes 3.5% to 5% of the ticket, not the 1% owners estimate from memory
In high-volume delivery operations — over 200 monthly orders per channel — packaging consumes between 3.5% and 5% of the sale ticket versus the 1% most owners budget from memory, or simply never budget at all. The gap comes from adding primary containers, outer bag, branded napkins, tamper-evident seal and, in cities with environmental regulations in force since 2024, the biodegradable packaging surcharge that in Mexico and Colombia can add $0.18-$0.35 USD per order. A restaurant with a $15 USD average ticket and 250 monthly orders that underestimates packaging by 3 percentage points loses $112.50 USD per month — over $1,350 USD per year — straight from gross margin, with no separate line item visible in any report if the cost system does not break it down by channel.
Invisible kitchen time: 3-6 extra minutes per order that nobody costs
Every delivery order draws between 3 and 6 additional minutes of kitchen labor for item verification, temperature double-check, packaging assembly and labeling — time that is never invoiced or included in any dish cost. On a shift processing 40 delivery orders over 8 hours, those 5 average extra minutes add up to 3.3 hours of kitchen work assigned to no cost line. At a labor cost of $3.80 USD per hour in mid-tier Latin American markets, that is $12.54 USD per shift or $376 USD per month in a restaurant running 30 delivery days. The secondary impact is equally relevant: that extra time delays dine-in dishes, creates slow tickets and triggers table complaints that erode Google review scores — which in turn undermines local organic search ranking.
How to build the delivery P&L: the four line items that cannot be missing?
An operational delivery income statement must segregate four line items that a dining-room P&L typically merges or ignores: (1) total effective app commission — base plus ads, expressed as a percentage of gross channel revenue;
(2) food cost recalculated using the delivery recipe, including sauces and add-ons; (3) packaging cost per order multiplied by monthly volume; and (4) additional kitchen time valued at hourly labor cost. Below 15%, the channel is destroying margin in absolute terms even if order volume grows; growing volume on a negative margin only accelerates the loss, it does not correct it.
App menu pricing: the equation Diego F. Parra applies in audits
Diego F. Parra applies a minimum delivery price formula in restaurant audits that starts from total dish cost — ingredients plus packaging plus valued kitchen time — and divides it by the complement of the effective commission: Min_price = Total_cost / (1 − Effective_commission). For a dish with a total cost of $6.20 USD and an effective commission of 30%, the minimum price is $8.86 USD; if the restaurant also targets a 20% delivery gross margin, the price rises to $11.07 USD. In practice, most operators set the app price equal to — or just 10% above — the dine-in price, which on dishes with a real commission of 33% — base plus ads — turns every sale into a net loss of 3 to 7 margin points. Channel-specific price adjustment is not optional: it is the first and most immediately actionable lever for delivery profitability.
Dark kitchens and traditional restaurants: delivery margins under the microscope in 2026
Dark kitchens that operate exclusively through apps, with 100% of sales in delivery, hold a structural advantage: they eliminate dining room costs — servers, tableware, décor, dining-area square footage — which in a traditional restaurant represent 18% to 25% of total operating expenses. However, some dark kitchen operations end up with negative net margins because the app commission plus sponsored advertising structure absorbs the fixed-cost advantage entirely. In traditional restaurants, profitable delivery requires an app price 15%-25% higher than the dine-in price to sustain the same gross margin; operators who skip that adjustment are unknowingly subsidizing the delivery channel with dining-room margin. In both models, the difference between running with or without a separate delivery P&L can represent 4-9 net margin points annually.
The 4 differences nobody costs out
Variable, not fixed, commission: the base rate of 25%-30% climbs another 3%-8% when the restaurant activates in-app sponsored ads, a cost almost nobody measures separately. Food cost without re-costing: a dish running 28% food cost in the dining room can hit 34%-37% in delivery if the portion, extra sauce and packaging aren't redesigned for the channel. Underestimated packaging: in high-volume operations (200+ monthly orders), packaging eats up 3.5%-5% of the ticket versus the 1% most owners budget from memory. Invisible kitchen time: every delivery order steals an extra 3 to 6 minutes of kitchen time for verification, double-checking and packaging, time that's never billed or measured in the dish's costing.
Myth vs reality: a direct, criterion-by-criterion analysis
The myth: delivery equals extra margin
- Net margin of +15%, identical to dine-in, with nothing adjusted.
- Fixed commission the channel has charged since that sales line opened.
- Packaging is almost free: barely 1% of the ticket, invisible in the costing.
- Kitchen time equal to dine-in: 8 minutes per order, with zero extra friction.
The reality: delivery needs its own unit economics
- Real net margin per order tends to swing between a small negative number and a low single-digit positive, far from what the platform dashboard shows.
- Effective commission of 25%-30% in 2026, peaking at 35% on orders with active sponsored ads.
- Real packaging cost of 3.5%-5% of the ticket in restaurants with 200+ monthly orders.
- Real prep time of 11-14 minutes due to order verification and extra packaging steps.
Delivery by the numbers: 2026
“We had a ceviche restaurant in Bogotá with 40% of sales coming from delivery, and the owner swore it was his most profitable channel. When we re-costed every dish with the real 28% commission, seafood-specific packaging and the actual assembly time, we found that 7 of his 12 top delivery dishes were losing between 2% and 9% net margin every single time they sold. We raised the delivery menu price 12% above dine-in, redesigned the packaging to cut its cost by 1.8 percentage points, and within 60 days the channel went from -3% to +8% real net margin, without losing order volume.”
Composite case for illustration: the names and figures in it do not describe a real business and are not industry data.
How to re-cost your delivery menu in 4 steps
Pull each app's statement (Uber Eats, Rappi, DoorDash) for the last 90 days and calculate the effective real commission, which usually runs 3-8 points higher than the advertised rate because of active sponsored ads the owner never notices.
Calculate channel-specific food cost by adding packaging, extra sauces and transport shrinkage; if the result tops the recommended 32% ceiling, adjust the portion or raise the delivery menu price 8%-15% above dine-in.
Time 20 delivery orders against 20 dine-in orders during peak hour; the 3-6 minute difference per order is real labor cost that almost no restaurant ever loads onto the channel's unit economics.
Subtract commission, packaging, adjusted food cost and extra time from each dish's sale price; if net margin falls below 5%, the channel needs a price adjustment, a shorter menu or renegotiated app terms.
And with AI?
Optimize channels, pricing and unit economics of your dark kitchen. Diego F. Parra is an expert in AI applied to restaurants.
Free tools for delivery unit economics
Tools to control delivery unit economics
Controlling unit economics by channel isn't done on a spreadsheet isolated from the rest of the business; it's done by connecting delivery sales, costs and cash flow to the same system you use for the whole restaurant, so no commission or packaging cost ever goes unmeasured.
Frequently asked questions about delivery unit economics
How do delivery fees affect restaurant margins at each location?
How do delivery fees affect restaurant margins at each location?
Delivery fees cut margin on every order, and the impact differs by location because ticket size, distance, platform mix and labor change from one store to another. With U.S. delivery orders averaging 20–35 USD (Lightspeed), a fixed fee or percentage commission weighs far more on a small ticket than on a large one, so the same menu can be profitable in one store and lose money in the next. The right unit is contribution per order per location: ticket minus food, packaging, commission and fees. Diego F. Parra's Masterestaurant method runs that calculation store by store before adjusting delivery prices, minimum order or delivery radius.
What's the real commission delivery apps charge in 2026?
What's the real commission delivery apps charge in 2026?
The base commission runs 25%-30% of the gross ticket on Uber Eats, Rappi and DoorDash, but once you add sponsored ads and card-payment fees, the effective real commission climbs to 30%-38% for most restaurants that haven't negotiated volume-based terms.
How do I know if my delivery food cost is too high?
How do I know if my delivery food cost is too high?
If the delivery dish's food cost tops 32% once packaging and transport shrinkage are added, the dish is in loss territory. The goal is keeping it at or below the dine-in food cost by adjusting price or portion per channel.
Is delivery always less profitable than dine-in?
Is delivery always less profitable than dine-in?
Not necessarily: with proper re-costing and a differentiated price 8%-15% above the dine-in menu, the delivery channel can reach 6%-10% real net margin. The problem isn't the channel itself, it's running it on dine-in prices and costs without adjusting anything.
How long does it take to re-cost a full delivery menu?
How long does it take to re-cost a full delivery menu?
A full re-cost of a 15-20 dish delivery menu takes 2 to 3 weeks: one week to audit real commissions and prep times, and two weeks to adjust prices, packaging and portions dish by dish based on the data.
Delivery unit economics by the numbers (2026)
Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.
| Metric | Value | Source |
|---|---|---|
| labor cost reduction achievable with AI-driven shift scheduling (8-12% range) | 8-12% (2023) | McKinsey & Company (cited in a TimeForge article): How AI-Driven Scheduling Cuts Restaurant Labor Costs 2023 |
| Cloud kitchen market by 2035 | USD 248.10 mil millones proyectados para 2035 | Precedence Research 2025 |
| Worldwide online food delivery revenue 2026 | USD 1.51 billones en 2026; CAGR 6.24% (2026-2031) | Statista 2026 |
| US online food delivery revenue 2026 | USD 473.49 mil millones en 2026 | Statista 2026 |
| Largest delivery market (China) 2026 | USD 539.87 billion in revenue in China in 2026 | Statista 2026 |
| Spain food delivery & dark kitchens market | Aprox. USD 5 mil millones | Ken Research 2025 |
Related content
Delivery unit economics with the Masterestaurant method
Applied in +8.400 restaurants across 43 countries.
