The Invisible Acquisition Cost: What Platforms Don't Tell Your P&L

1. Why isn't the 30% commission the cost of delivery?

The 30% commission is not the cost of delivery: it's the price of renting a customer who will never be yours. I've seen it across dozens of operations we audit at Masterestaurant:

the owner reads the P&L by gross sales line, celebrates that delivery is 40% of total tickets, and never separates how much of that revenue pays acquisition versus how much leaves margin. On a 20 USD ticket, the platform keeps 6 USD in commission, add 2 USD in promo discount and 0.80 USD in repackaging: 8.80 USD gone before touching food cost. If your plate cost already starts at 30%, the app order leaves a negative contribution margin. Diego F. Parra sums it up in one boardroom line: the channel that bills the most is the one that keeps the least, and nobody audits it. The problem is accounting before it's operational: when 40% of your sales come from a third party that owns the diner relationship, you pay a recurring customer acquisition cost for people who never order from you directly again.

2. The problem is accounting before it's operational

The traditional model logs that order as clean revenue; the Masterestaurant approach reclassifies it as CAC that never amortizes, because the same diner triggers a fresh 30% commission on every reorder. In healthy marketing, CAC is paid once and LTV recovers it over months. Not here: the platform charges the acquisition toll on order one, order five, and order twenty. On 100,000 USD in annual delivery, that's 30,000 USD leaking into customers the app resells to you every single month. Separating CAC from contribution margin is what turns a volume celebration into an EBITDA decision. The traditional model blends the two and concludes delivery is "growing"; the hybrid architecture splits them and only then measures whether the channel adds or subtracts. Do it in three lines: channel gross sales, minus commission and discounts (the real CAC), minus variable product and repackaging cost. What remains is true contribution.

3. Separating CAC from contribution margin changes the decision

In the operations we reviewed, a channel billing 12,000 USD monthly left 400 USD in contribution after the toll: 3.3%, when the dining room left 22%. The conclusion isn't to kill delivery, it's to stop subsidizing it with dining-room cash. A channel that subtracts from EBITDA isn't volume, it's a hemorrhage disguised as growth. Capturing the diner on the first order is what breaks the platform's eternal toll. In the status quo, the app owns the relationship and the data; in a well-designed hybrid model, the restaurant uses the platform as an entry door and moves the reorder to a channel with near-zero commission. The mechanism is concrete: every delivery order ships with a physical insert offering an incentive to order direct next time —a 15% that still costs you less than the 30% commission— and a QR that loads the diner's WhatsApp number into your own database.

4. Capturing the diner on the first order

I've seen operations migrate from 8% to 35% direct reorder in four months. Every order that shifts from the app to your direct channel recovers 5-6 USD per ticket you used to hand the intermediary, and that diner is now yours. Online reputation stops being scattered noise across three apps once it's managed as a measurable conversion lever. In the traditional model, reviews land ownerless on each platform and nobody responds; in the hybrid architecture, a sub-24-hour response protocol and a request routine after every well-resolved order raise the rating 0.6 stars in six months. That's not cosmetic: in delivery, half a star moves your position in the algorithm's ranking and the funnel conversion rate. In the operations we support, going from 4.1 to 4.7 stars lifted impression-to-order conversion from 2.8% to 4.1% —46% more orders with the same traffic—.

5. Online reputation as a managed conversion lever

Managed reputation is the only delivery-channel lever whose return the platform commission doesn't get to keep. Platform delivery does pay off when you treat it as customer-acquisition marketing spend, not as a profitable revenue line on its own. The math is cold: if each new customer the app brings and you manage to migrate to direct has a 240 USD one-year LTV —12 orders of 20 USD—, paying 8 USD commission on the first order is a 3.3% CAC on that lifetime value: excellent. The mistake is paying that same CAC 12 times because you never captured the diner. The Masterestaurant rule is simple: an app order is justified only if your direct-channel capture rate exceeds 30%; below that, you're renting customers at a dead loss. Measure monthly capture, not gross sales: it's the only number that tells you whether the channel builds your business or the platform's.

6. The rebuilt P&L: what to watch every week

The rebuilt P&L separates four lines the platform report hides on purpose: channel gross sales, commission and discounts as CAC, variable product cost, and direct-channel capture rate. Watch them every Monday, not every quarter. The report the app sends bundles everything into "your earnings" so you celebrate volume and never audit the toll; Diego F. Parra insists that number is the platform's marketing, not your accounting. An 80,000 USD monthly operation that audits this way spots patterns fast: which dishes lose in delivery to repackaging, which time slot brings capturable customers and which only brings discount hunters. Rebuilding the P&L costs a spreadsheet and two hours a month; ignoring it costs 30% of your biggest channel, every month, in customers who will never be yours. The traditional model confuses gross delivery revenue with profitability; the Masterestaurant approach separates customer acquisition cost from contribution margin and only then measures whether the channel adds to or subtracts from EBITDA.

7. The differences that decide channel EBITDA

In the status quo the platform owns the relationship and the data; in the hybrid architecture the restaurant captures the diner on the first order and shifts repeat purchases to a channel with near-zero commission. Online reputation moves from scattered noise across three apps to a managed conversion lever that rises 0.6 stars in six months and lifts the delivery conversion rate of the sales funnel.