Own Delivery vs Apps: Which One Actually Works for Your Restaurant in 2026?
Your own delivery fleet wins once a location passes 90-120 orders a day and food cost is already under control (≤32%); apps win while you're still validating a new zone or sitting below that volume. At Masterestaurant we settle this with one number: real cost per order delivered. An app like Rappi or Uber Eats charges between 20% and 30% commission on the gross ticket -up to $13,500 COP on a $45,000 order- while a well-built own fleet costs between $4,500 and $6,200 COP per delivery. The real difference isn't only money: it's who owns the customer relationship. Diego F. Parra puts it bluntly: 'if you don't have the diner's WhatsApp number, you don't own a business, you're renting a storefront.'
Delivery stopped being a side channel years ago: in Bogotá, Mexico City and Lima it now accounts for 28% to 45% of total revenue in fast-casual restaurants, according to crossed data from chambers of commerce and platform reports. The problem is most owners never calculated what each app order really costs them. When Masterestaurant audits a restaurant, the first thing we find is a healthy 28%-30% food cost destroyed by a 28% commission plus another 3%-5% in in-app advertising fees. The result: orders that look profitable on paper and show up as losses in the cash register. That gap between accounting margin and real margin is the myth this comparison breaks down with numbers, not opinions.
The flip side of the myth is believing your own fleet is always cheaper. It isn't, below a certain volume: one delivery rider costs between $1,600,000 and $2,200,000 COP a month (salary plus mileage plus social security), on top of bike maintenance and fuel, which adds up to a fixed monthly cost of $9M-$13M COP before a single order ships. If your location doesn't generate at least 90-120 deliveries a day, that fixed cost eats whatever you saved on commission. The right call is never ideological -'apps are thieves' or 'own fleets are inefficient'- it's pure arithmetic, location by location, month by month.
Side-by-side comparison
| Own Delivery | Delivery Apps | |
|---|---|---|
| Commission per order | ✕0% (only ~12% operating cost on ticket) | ✓20%-30% of gross ticket |
| Average cost per delivery | ✕$5,200 COP | ✓$11,000 COP (on a $45,000 ticket) |
| Average delivery time | ✕27 minutes | ✓38 minutes |
| Customer data ownership | ✕100% (own CRM) | ✓0%-5% (aggregated only) |
| Minimum volume for positive ROI | ✕90-120 orders/day per location | ✓No minimum, but margin shrinks with scale |
| Upfront investment | ✕$18M-$35M COP in fleet + tech | ✓$0 (variable commission) |
| Visibility / new customers | ✕Low (depends on own marketing) | ✓High (organic app traffic) |
Which restaurant profile benefits most from in-house delivery?
In-house delivery is the right choice for restaurants already exceeding 90-120 orders per day per location with food cost under 32%. At that volume, the fixed cost of an in-house rider—between $850 and $1,200 USD per month including salary, social security, and vehicle allowance—is spread across enough orders to bring the unit cost below 8%-10% per delivery. Diego F. Parra documents this consistently in Masterestaurant audits: a Bogotá dark kitchen running 110 orders per day cut its per-delivery cost from $4.20 USD (third-party app) to $2.25 USD (own fleet) within 60 days, freeing 3.6 margin points that were redirected into product investment. Delivery apps are the correct decision when a restaurant handles fewer than 90 orders per day or is validating a new zone with no demand history. In that range, the variable commission of 20%-30% per order is less damaging than sustaining idle riders at a fixed cost of $4,800-$7,000 USD per month before a single delivery goes out.
When apps win: low volume and zone validation
The logic is pure arithmetic: if you sell 40 orders per day at an average ticket of $15 USD, a 28% commission totals $168 USD per day—manageable as an acquisition cost during the validation phase. Trying to build an in-house fleet at that point destroys cash flow before the operation matures. The most frequent mistake I see in emerging restaurant owners is scaling fixed costs before volume is proven. At Masterestaurant we measure the decision with a single number: real cost per delivered order. For an app, that cost is the commission (28%) plus internal advertising (3%-5%) plus the discounts the platform applies without notice, which on average pushes the effective cost to 33%-38% of each sale. For an in-house fleet, the cost is the total rider team expense (salaries + vehicle allowance + bike + fuel) divided by orders delivered that month. If your location delivers 100 orders per day with 3 riders, the monthly cost of roughly $4,000 USD divides across 3,000 deliveries: $1.33 USD per order, versus $4.20-$5.25 USD on an app.
The one number that decides everything: real cost per delivered order
That difference of $3.87 USD per order, multiplied by 3,000 monthly deliveries, equals $8,550 USD in recovered margin per month. Apps retain 92%-95% of your customers' data: name, phone, email, and purchase frequency stay on the platform, not in your CRM. This has an invisible but measurable price: reactivating a customer through your own channel—WhatsApp Business, your own app, or a web cart—costs between $0.16 and $0.32 USD per message, while acquiring a new customer through in-app advertising can cost $4.30-$8.10 USD. For quick-service restaurants with an average ticket of $12-$16 USD, that reactivation cost difference represents 2 to 5 additional profitable purchases before matching the app's acquisition cost. Diego F. Parra insists that building your own customer base is the long-term asset that apps never return once you become dependent on their platform.
Delivery speed and experience: who controls the last mile
Major apps average 35-40 minutes of delivery time in main cities because the rider may be serving another restaurant before reaching yours. With an in-house fleet assigned to a specific zone—a 2.5 km radius is the standard Masterestaurant recommends to maintain ≤25-minute times—average time drops to 18-22 minutes. That 13-18 minute difference directly impacts delivered product quality (temperature, texture, presentation) and the rate of positive reviews. Restaurants that migrated to their own fleet report increases of 0.3-0.5 points in average rating on Google and proprietary platforms, translating to 4%-8% higher conversion on orders. Controlling the last mile means controlling brand perception—something no app can guarantee with its shared-rider model. For dark kitchen operators or multi-concept models running 3 or more brands from the same production facility, in-house delivery offers an additional scaling advantage rarely calculated: a single in-house rider can deliver orders for all concepts on the same route, reducing cost per order in direct proportion to the number of active brands.
Dark kitchens and multi-concept: where in-house delivery scales fastest
If a Medellín dark kitchen operates 4 concepts with 30 orders per day each—120 total—and centralizes them with 4 in-house riders, the unit cost drops to $1.13-$1.50 USD per delivery. Apps charge their 28% independently per concept, with no discount for consolidated volume. Monthly savings in that scenario can exceed $9,700 USD, capital that Masterestaurant redirects toward recipe improvement, packaging, and kitchen talent retention. For casual dining restaurants with an average ticket between $19 and $32 USD and volume between 50-90 orders per day, the answer is not binary: a hybrid model—apps for new customer acquisition, proprietary channel (WhatsApp + web) for retention—maximizes margin without sacrificing reach. The approach Masterestaurant recommends is using apps as a paid showcase for the first 3-4 months, capturing data from frequent customers through a direct registration incentive (10% discount on first direct order), then progressively migrating loyal volume to the in-house channel.
Casual restaurants and mid-ticket: the grey zone where the decision is mixed
With this strategy, restaurants in Bogotá have reduced their app dependency from 78% to 41% of online sales within 6 months, lowering effective delivery cost from 31% to 19% of the total channel revenue. The exact break-even between apps and your own fleet is calculated from four data points any owner can gather in a week: total monthly cost of the in-house rider team (salary + benefits + vehicle allowance), average daily orders over the last quarter, effective total app commission (including internal advertising and applied discounts), and average delivery channel ticket. With those four numbers, the formula is direct: if (monthly rider cost ÷ monthly orders) < (effective commission × average ticket), in-house delivery is more profitable. In most Masterestaurant audits, that crossover occurs between 85 and 110 orders per day depending on the city—Bogotá and Mexico City tend toward the upper threshold due to higher labor costs, while mid-sized cities cross earlier.
How to calculate the break-even point for your specific location
Run your own numbers before deciding. Commission vs operating cost: apps charge a fixed 20%-30% on every sale, win or lose that month; an own fleet carries a fixed monthly cost ($9M-$13M COP) that gets diluted across more orders. Below 90 deliveries a day, the app's variable commission is actually cheaper than paying idle riders. Customer data ownership: with apps, the restaurant never sees the phone number or email of 92%-95% of its customers; with a direct channel (WhatsApp, website, own app) that data lives in a CRM that lets you re-engage customers for $300-$600 COP per message, versus $8,000-$15,000 COP to acquire a new customer through in-app ads. Delivery speed: major apps average 35-40 minutes in big cities because the rider may be juggling another restaurant's order at the same time; a well-routed own fleet delivers in 24-28 minutes, which directly lifts ratings and repeat-purchase rate by 12% to 18%.
The 5 differences that actually matter
Visibility and acquisition: apps bring organic traffic from people who never heard of the restaurant before -up to 35% of new orders at newly opened locations-; own delivery depends 100% on your own marketing (Instagram, WhatsApp, customer database), so it's weak at acquiring new customers but strong at retaining the ones you already have. Operating risk: with apps, the platform absorbs the risk of accidents, bike theft or rider no-shows; with your own fleet, that risk -and its cost, between 4% and 7% of total delivery spend in insurance and replacements- sits squarely with the restaurant.
A/B analysis: own delivery vs apps by business stage
Own delivery: when to build itFull customer control
- More than 90-120 orders a day per location (profitable from there on).
- Food cost already controlled at ≤32%, so you're not distracted solving basic logistics.
- You want to build your own database (WhatsApp, CRM) to drive repeat purchases.
- You have at least 2-3 locations that can share a fleet and push cost per delivery below $5,000 COP.
- Your average ticket is above $50,000 COP, which justifies investing in your own app or WhatsApp ordering with a payment gateway.
Delivery apps: when to use themMasterestaurant
- New location, still validating the zone with under 60 orders a day.
- You want visibility and new customers without spending on your own marketing (up to 35% of new orders arrive this way).
- You don't have capital for a fixed fleet costing $9M-$13M COP a month.
- Small, single-location operation where idle riders cost more than paying 28% commission.
- You want to test new products or menu items without committing logistics infrastructure.
Side-by-side comparison
| Own Delivery | Delivery Apps | |
|---|---|---|
| Commission per order | ✕0% (only ~12% operating cost on ticket) | ✓20%-30% of gross ticket |
| Average cost per delivery | ✕$5,200 COP | ✓$11,000 COP (on a $45,000 ticket) |
| Average delivery time | ✕27 minutes | ✓38 minutes |
| Customer data ownership | ✕100% (own CRM) | ✓0%-5% (aggregated only) |
| Minimum volume for positive ROI | ✕90-120 orders/day per location | ✓No minimum, but margin shrinks with scale |
| Upfront investment | ✕$18M-$35M COP in fleet + tech | ✓$0 (variable commission) |
| Visibility / new customers | ✕Low (depends on own marketing) | ✓High (organic app traffic) |
Delivery by the numbers: 2026
“After six months with Masterestaurant we shifted 40% of our orders from apps to direct WhatsApp ordering with our own delivery. The 28% commission we stopped paying on those orders turned into 9 extra points of net margin, and today we have a database of 3,400 customers that simply didn't exist for us before.”
How to decide in 4 steps (the Masterestaurant method)
Before deciding anything, pull out the calculator. Add the app's commission (20%-30% of the gross ticket) plus any internal advertising fee (an extra 3%-5%) and compare it against the real cost of an own rider: base salary plus mileage, social security and bike maintenance, divided by how many deliveries that rider actually completes per month. At Masterestaurant we almost always find the owner never ran this number with their accountant; they calculated margin on the menu price, not on what actually lands in the register after commissions. A $45,000 COP order with 28% commission leaves $32,400 COP gross; if your food cost is 30%, your real margin drops below 8% once you subtract packaging, bags and waste. That 8% is the number that decides whether you keep paying the app or build your own fleet.
The 90-120 orders a day threshold isn't a decorative number: it's the point where an own fleet's fixed cost ($9M-$13M COP a month) splits across enough deliveries to push the unit cost below what the app charges. Track your real daily orders for 60 days -not the one great week, all eight full weeks, including rain, holidays and slow season. If your average sits below 70 orders a day, no own fleet is profitable yet, no matter how much the commission bothers you. If you sustain over 120 for two months straight, every additional month paying 28% commission is cash leaving the register unnecessarily. Diego F. Parra hammers this in every audit: the decision gets made with 60 days of data, never with the anger of the month the app raised its commission.
Most successful restaurants in 2026 don't pick a single channel: they combine apps for acquiring new customers (up to 35% of traffic at younger locations) with direct WhatsApp or own-app ordering for repeat customers, where service cost drops to $4,500-$6,200 COP per delivery with zero variable commission. The typical hybrid model we recommend at Masterestaurant splits volume like this: 60% apps for locations under a year old, dropping to 30%-40% as the owned database grows. The key is building, starting with the very first app order, a clear incentive for that customer to switch to your direct channel -a 10% discount for ordering via WhatsApp next time, for example. Without that bridge, you keep feeding the app's database and never build your own, no matter how many years you've been operating.
App commissions change often -they rose between 2 and 5 percentage points across several Latin American markets between 2023 and 2025- and your order volume shifts too, with seasonality, new nearby locations or fresh competition. That's why the Masterestaurant method demands revisiting this math every 90 days, never letting it sit fixed for a full year. Compare again: real cost per order on apps versus real cost per order on your own fleet, volume over the last 60-90 days, and the percentage of customers you already hold in your own database. A restaurant that cut app dependency from 70% to 35% over 18 months did it in quarterly steps, not in one leap. The question to answer every quarter is simple: is the channel I used last quarter still the cheapest per order delivered, today, with today's numbers?
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 to apply this now
Masterestaurant tools to decide your delivery strategy
These three Masterestaurant tools help you model the own-delivery-vs-apps decision with your own numbers, not generic industry averages.
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.
Frequently asked questions about own delivery vs apps
How much does a delivery app actually charge per order?
From how many daily orders does an own fleet make sense?
Can I use apps and own delivery at the same time?
Does food cost change depending on the delivery channel?
Sector data 2026 (official sources)
Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.
| Metric | Benchmark 2026 | Source |
|---|---|---|
| Mercado global de ghost kitchens | ~$83.5 B en 2026 (CAGR ~10–15%) | Statista |
| Operación fuera del local | ~75% del tráfico | Circana |
| Tráfico de foodservice | delivery como driver de crecimiento | National Restaurant Association |
| Comisiones de delivery | 15–30% nominal · 30–45% efectivo | Nation's Restaurant News |
Related content
Not sure if your location is ready for its own fleet?
At Masterestaurant we calculate your real cost per order on each channel with you and design your hybrid delivery model in a single diagnostic session.