Own delivery vs app commissions: the hidden cost mistakes destroying your margin and the right method 2026

In 2026, delivery represents 18%–35% of total sales for urban restaurants in Mexico and Latin America, according to CANIRAC data and regional operator reports.

The three dominant apps — Rappi, Uber Eats, and DiDi Food — charge between 25% and 35% commission on the consumer sale price, plus additional platform advertising fees (boost) ranging from 5%–15% extra for relevant positioning.

The most common mistake Diego F. Parra observes in restaurant owners is comparing own delivery vs apps by looking only at the commission percentage, ignoring hidden costs of the own channel: delivery staff salary and benefits, motorcycle insurance, maintenance, fuel, coordination time, tracking technology, and the cost of failed or late orders that trigger refunds.

Masterestaurant has audited delivery cost structures in over 60 restaurants across Mexico, Colombia, and Peru between 2023 and 2026. The consistent finding: 74% of restaurants operating their own delivery underestimate their real cost per order by at least 40%.

What apps really charge: base commission plus hidden costs

The three dominant apps in Mexico and Latin America — Rappi, Uber Eats, and DiDi Food — charge between 25% and 35% commission on the consumer sale price, but that percentage is only the floor. Restaurants that invest in in-platform visibility pay an additional 5% to 15% in boost campaigns to appear in relevant positions; without that spend, conversion drops 30% to 50% compared to establishments that do invest. Add the premium packaging required by platforms — sealed bags, security stickers, insulated containers — which adds between $0.40 and $0.75 USD per order. On an average ticket of $9 USD, a 30% base commission is $2.70; add an 8% boost and $0.60 packaging, and the real charge reaches $4.02 per order, equivalent to 44.7% of the ticket. At that level, any restaurant with a 28% food cost operates at a net loss on every app delivery. In-house delivery is not inherently cheaper — it is only cost-effective when volume justifies the structure.

The real cost of in-house delivery: per-order math that few operators run

A delivery driver in Mexico City with base salary, social security contributions, motorcycle use, and fuel represents a monthly cost of $600 to $900 USD. At 60 daily orders (1,800 per month), the per-order cost for the driver alone is $0.33 to $0.50 USD. But at 30 orders per day that cost rises to $0.65–$1.00 USD per order for that line item alone, before accounting for tracking technology ($25–$60 USD/month), internal coordination ($100–$200 USD/month in staff time), or failed and late orders that generate partial refunds. The real efficiency threshold for the in-house channel, according to Masterestaurant's audit of over 60 restaurants in Mexico, Colombia, and Peru (2023–2026), is ≥80 orders per day sustained; below that, apps with a negotiated commission are lower in total variable cost per order. Diego F.

74% of restaurants underestimate their in-house delivery cost by at least 40%

Parra and the Masterestaurant team have audited delivery cost structures in more than 60 restaurants across Mexico, Colombia, and Peru between 2023 and 2026, and the finding is consistent: 74% of operators running their own fleet underestimate their real per-order cost by at least 40%. The most common mistake is calculating only the driver's salary and fuel, ignoring statutory benefits — vacation pay, year-end bonus, social security — which increase gross labor cost by 35% to 42% above net wages, plus motorcycle maintenance ($75–$150 USD/month under intensive use), liability insurance ($40–$90 USD/month), and equipment depreciation. When all those line items are loaded, the real per-order cost on an in-house channel running 40 deliveries per day ranges from $1.10 to $1.90 USD — a figure that, compared honestly against a 28% app commission on a $9 ticket ($2.52), narrows the advantage of the in-house channel significantly.

How apps distort price and erode purchase frequency

To absorb part of their commission without sacrificing margin, most restaurants inflate in-app prices by 10% to 20% versus the dine-in menu. That gap is now visible to consumers: in 2025, Uber Eats and Rappi enabled explicit notices when in-app prices differ from in-store prices, creating trust friction. The operational effect is a decline in recurring order frequency: customers who order weekly tend to drop to every 12–18 days when the in-app ticket is 15% higher than in-store. In higher average-ticket categories ($17–$30 USD), that differential reduces perceived value and increases cart abandonment rates by 18% to 27%, according to internal data from operators processing more than 500 orders per month on platforms. Masterestaurant's recommendation is price parity with margins adjusted from the cost structure — not menu inflation — to preserve customer lifetime frequency. The breakeven analysis between in-house delivery and apps depends on three variables: daily order volume, average ticket, and coverage zone.

Financial breakeven: when each channel makes sense by volume

With a $11 USD average ticket and a 30% app commission ($3.30 per order), in-house delivery becomes more economical when its per-order cost falls below $3.30. According to Masterestaurant audit data, that occurs at 75–85 orders per day in an urban zone with a maximum 3 km radius. At 50 orders per day, the in-house channel carries a cost of $1.40–$2.25 per order (excluding technology), still below the app commission, but the gap narrows to 20%–35% — and disappears if one in every 12 orders fails, because that refund absorbs the cumulative savings from the previous 11 successful orders. The real threshold is not fixed: each restaurant must run it with its own labor cost mix, zone, and actual volume, not sector averages. Total dependence on apps creates a concrete business risk: when Rappi or Uber Eats change their algorithm or raise commissions — which happened in Mexico in 2024, when several platforms raised their base rate by 2 to 4 percentage points without prior notice — the restaurant has no negotiating leverage.

Hybrid structure: the model that breaks the full-dependency trap

The model Masterestaurant recommends after analyzing more than 60 operations is a hybrid scheme: maintain app presence to capture new demand (customers who don't know the restaurant yet) with a minimum ticket of $9 USD to ensure positive margin after commission, and build an in-house channel for the recurring customer base. Restaurants that migrate 35% to 45% of their recurring orders to the in-house channel (WhatsApp plus proprietary ordering system) report a margin improvement per order of 12 to 19 percentage points in that segment, while maintaining total volume through apps. Many restaurant owners assume app commission is fixed, but in practice three negotiation levers exist for operators with sufficient volume. First: volume-tiered commission — Rappi and Uber Eats have private schemes that reduce the commission by up to 3–5 percentage points when the restaurant exceeds a monthly order threshold (typically 300–500 orders per month).

What to negotiate with apps before assuming there is no way out

Second: temporary exclusivity — some platforms offer a 2%–4% commission reduction in exchange for exclusivity over 3–6 months; weigh the opportunity cost carefully before signing. Third: bundled visibility packages — instead of paying additional boost fees, negotiate visibility campaigns as part of the base contract. The most common negotiation mistake Masterestaurant documents: the owner negotiates the commission rate without first measuring their real per-order cost on the in-house channel, so they accept terms without knowing whether those terms are favorable compared to their actual alternative. The in-house vs app delivery decision is not ideological — it is arithmetic. The first step is building a real per-order cost sheet for each channel with every line item loaded: for in-house, add gross driver cost (including benefits), fuel, maintenance, insurance, technology, and internal coordination; divide by actual orders from last month, not projected volume. For the app channel, add the commission percentage on your real average ticket, premium packaging cost, and monthly boost or advertising spend averaged per order.

Concrete action: audit your per-order cost this week, not next year

With those two real figures on the table, the breakeven point is self-evident. In the restaurants audited by Masterestaurant, this exercise takes under 90 minutes, and in 68% of cases reveals that the optimal channel mix differs from the current one — meaning margin being left on the table today that can be recovered in the next billing cycle. App commission is not the only cost: add the premium packaging required by platforms (sealed bags, security stickers) at $0.40–$0.80 USD per order, plus the staff time to prepare and hand off to the app's delivery driver. Own delivery is not 'cheaper' by default. A delivery rider in a major Latin American city with salary, social security, motorcycle, and fuel costs $600–$900 USD/month. At 60 orders/day (1,800/month) that is $0.33–$0.50 USD per order. At 30 orders/day the cost doubles to $0.67–$1.00 USD per order for the rider alone.

Key differences most restaurant owners ignore

Apps inflate consumer prices 10%–20% to partially absorb their commission. This reduces repeat purchase frequency: a weekly dine-in customer drops to every 2–3 weeks via app due to the higher perceived price. With apps, the restaurant loses the direct customer relationship. Without its own data, retention cost becomes invisible but permanent: every repurchase passes through the platform and pays commission again, while own delivery amortizes the acquisition cost across subsequent orders at zero additional commission. Rejections and refunds in own delivery fall 100% on the restaurant. Apps can recover 60%–70% of rejected order value depending on the contract, but the administrative time per incident represents an unquantified hidden cost that Diego F. Parra estimates at $0.75–$1.25 USD per incident. The hybrid model — apps for customer acquisition + WhatsApp/own app for retention — is the strategy Masterestaurant has implemented with the best results: it reduces effective commission to 18%–22% by migrating repeat customers to the direct channel after their first app order.

A/B Analysis: Own Delivery vs Third-Party Apps — criterion by criterion

Hard delivery cost data 2026

4 steps to calculate which delivery channel is right for your restaurant

Free tools to apply this now

Masterestaurant tools to control delivery costs

Calculating delivery cost without tools is like driving without a speedometer: you get there, but without knowing how fast you are running out of fuel. These three Masterestaurant ecosystem tools work together to give you the exact number, not an estimate.

The right sequence: first structure your business model with Canvas, then project the break-even point by channel with Exponencial, and monitor in real time with Cash. The combination eliminates 80% of the assumptions currently destroying margin in your delivery operation.

Diego F. Parra — International consultant, expert in creating and scaling restaurants and in AI applied to restaurants, foodtech and HORECA. Methodology applied in 8.400+ restaurants across 43 countries · Expert in Artificial Intelligence applied to restaurants, hospitality and food businesses · 20+ years in restaurants, catering, large events and business growth · Author of the book «From Slave to Owner» (Amazon) · International keynote speaker for the HORECA sector.

Full profile →

FAQ: own delivery cost vs app commissions

Sector data 2026 (official sources)

Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.

Related content