Unit economics of delivery: checklist of mistakes vs the right method

What is delivery unit economics and why does almost everyone calculate it wrong?

Delivery unit economics is the real margin left by ONE order after subtracting commission, food cost, packaging, dispatch and bounced orders — not the revenue figure the app displays.

The owner sees Rappi deposit USD 12 per order and assumes a USD 12 profit; after commission (18-28%), packaging, the dish itself, the delay refund and the subsidized delivery fee, that order nets closer to USD 2.50. Diego F. Parra confirms this pattern auditing more than 800 restaurants a year at Masterestaurant: the leak isn't the platform, it's the absence of an 8-variable audit every 7 days. Nearly 75% of US restaurant traffic is already off-premise (National Restaurant Association 2025), so this calculation stopped being optional. Confusing gross revenue with net margin is the error behind menus that sell heavily and profit barely, and no promotion fixes it if the underlying math was wrong from the ticket up.

Real commission paid: the first line almost nobody breaks down

The compliance criterion is simple: if you don't know the exact commission percentage charged on the last 50 orders per aggregator, this line fails. Rappi and Uber Eats advertise one rate, but taxes, insurance and dynamic hour-or-zone adjustments push the real commission 2-3 points above the advertised figure — a hidden cost that erodes 3 to 5 margin points without ever showing up in a weekly report. New York already capped commissions at 15% per delivery and 5% for other services (Restaurant Business 2023), a signal that the market knows uncapped commission is unsustainable. Auditing every 50 orders, pulling the itemized breakdown instead of the net deposit, is the only way to see the true number. Anyone trusting the app's monthly summary is accepting the figure that favors the aggregator, not the one that describes their own cash register. The criterion: calculate food cost against the price the guest ACTUALLY paid, not the menu price, because 40% of Rappi orders use a coupon or aggregator promotion.

Food cost net of promotion: why the margin hides inside the discount

That discount isn't subtracted from the restaurant's food cost, only from the sale price, so margin compresses 8-12 points on every promoted order; if those orders make up 35% of volume — typical on Rappi — average margin drops 3 points automatically, quietly, month after month. This is where I got it wrong for years in early audits: I accepted food cost calculated against menu price and celebrated a 30% that, in the delivery channel, actually ran closer to 35-38%. The fix requires splitting food cost by channel — dine-in, delivery with coupon, delivery without coupon — instead of averaging them, because averaging is exactly where the early warning gets lost. The compliance criterion: measure the percentage of orders canceled for lack of an available courier against total weekly volume; between 6% and 9% of orders bounce for this reason, and each one carries two paths, both bad.

Zone bounce: the order you already cooked that nobody picked up

If the ingredient is already prepped and can't be repurposed, that order's margin is -100% because the cost is sunk with zero revenue; if the aggregator issues a credit, the administrative process eats hours from a team already running on thin margin. What Diego asks first when auditing dark kitchens is how many bounces are being actively reclaimed versus how many are quietly written off, because most restaurants never claim the credit they're owed. Without a daily log of bounces by zone and time slot, the business can't tell a demand problem from an aggregator courier-coverage problem, and ends up cutting prices when the real issue is logistics. Auditing this line takes ten minutes if the POS exports cancellations with a reason code; the blind spot isn't technical, it's a habit gap. Five failures repeat across Masterestaurant audits, and each carries a price tag in dollars, not theory.

The top 5 mistakes almost everyone makes in their delivery audit

First, blending dine-in and delivery food cost: it hides 3-5 margin points. Second, not breaking down the aggregator's real commission: another 2-3 points nobody sees in the monthly summary. Third, not reclaiming zone bounces: already-cooked orders lost at 100%, not 50%. Fourth, ignoring packaging as its own cost line: it averages 4-6% of ticket value and rarely appears in the dish's cost sheet. Fifth, not measuring assembly time per station: once it passes 6-8 minutes, the app penalizes ranking and organic volume drops, a silent punishment nobody connects back to the register. Combined, these five failures explain why a restaurant with 35-45% of sales through delivery — the tier-1 Latin American city average — can be running negative real margin while the gross revenue report looks perfectly healthy. Implementation needs a fixed owner, a fixed day and eight fixed numbers: the shift lead — general manager or kitchen chief, never the owner in small operations — runs the audit every Monday, cross-checking the aggregator report against the POS and waste log.

How do you implement this checklist into the restaurant's real routine?

The routine takes 30-40 minutes once the eight lines — commission, net food cost, assembly time, packaging, dispatch, bounce rate, discount and break-even margin — live in a fixed template instead of being reinvented weekly.

The most common implementation mistake is delegating it to someone without authority to fix anything: an audit with no decision power is a dead report. Masterestaurant pieces that apply this checklist see an 18-22% increase in net margin within 90 days, not from cutting prices but from closing specific leaks — misclassified commission, uncounted discounts, unclaimed bounces. Weekly frequency matters more than monthly depth because the aggregator's algorithm shifts conditions by time slot, and a monthly review arrives too late to fix the week already in progress.

How do you audit compliance on each item with measurable evidence?

Each of the eight lines needs exportable evidence, not a manager's impression: real commission is audited against the itemized breakdown of the last 50 orders, never the monthly summary;

net food cost is audited by splitting dine-in from delivery-with-coupon and delivery-without-coupon; assembly time is audited using the POS timestamp between order confirmation and courier handoff; packaging is audited by unit cost multiplied by actual monthly volume, not the cost sheet's estimate; bounce rate is audited against the cancellation log with an explicit reason code. Without this evidence, the audit is an opinion wearing a process costume. Break-even margin — the number that closes the checklist — is calculated by subtracting the previous seven lines from the sale price and comparing it against the recommended maximum food cost of 32%; below that threshold with all seven leaks closed, the delivery order leaves real margin, not just recorded revenue.

How do you audit compliance on each item with measurable evidence — in practice?

Masterestaurant carries this framework into restaurantecercademi audits because each aggregator's geography and algorithm belong to the local operation, never to the brand or to a generic template.

Hidden commission: Rappi and Uber embed tax, insurance, and dynamic zone/time adjustments in the % you see. Real commission is 18-28%; auditing by order every 50 sales uncovers 2-3% you didn't know existed. Margin impact: −3 to −5 points. Discount leakage: 40% of Rappi orders pay less because they used an aggregator coupon or promo. Not deducted from your food cost, only from sale price; margin compresses 8-12 points per discounted order. If 35% of volume is discounted (typical), average margin drops 3 points automatically. Zone rebate: 6-9% of orders cancel because driver is unavailable. Those 'unpaid' orders create ingredient waste (−100% margin) or admin rework (negative margin). Weekly zone audit reduces rebate to <3%. Packaging invisible: box + bag + label + napkin + utensil sum USD 0.45-0.65 per order, but many restaurants pass it as 'overhead.' 300 delivery orders/week = USD 135-195/week unassigned.

Top 5 Leaks Most Miss

Adding it to margin formula adjusts price or cuts electricity spend. Unknown break-even: You price assuming 20% margin, but real numbers (after all lines) yield 12%. Selling below break-even is pure loss. Margin matrix by SKU is what separates scale from volume-only churn.