The restaurant sells well on the delivery app yet still ends the month struggling. The explanation lies in the commission, which takes from 12% to 30% of each order, and something worse: the customer belongs to the app, not the restaurant. You don’t have their phone number, don’t know what they ordered last time, and can’t call them back on a slow Tuesday.
In this article
The commission calculation
A restaurant with R$ 90,000 in monthly delivery revenue, paying an average commission of 22%: that’s R$ 19,800 per month, or R$ 237,600 per year, leaving the bottom line.
Migrating just 30% of that volume to direct ordering represents R$ 27,000 monthly outside of commission fees. After deducting the real costs of this operation — own or third-party delivery per run, payment fees, service time — the net savings usually range between R$ 3,500 and R$ 5,500 per month. In a sector where net margins rarely exceed 8%, this equates to the profit from R$ 50,000 to R$ 68,000 in additional sales.
And there is an asset not reflected in the calculation: the customer base. Three hundred recurring customers with phone numbers, order history, and known frequency are worth more in the medium term than commission savings — because they allow you to generate demand instead of waiting for it.
Why you shouldn’t leave the app
It’s important to say clearly, because the above calculation often leads to a hasty decision: leaving the marketplace all at once is almost always a mistake.
The app delivers something the restaurant alone does not have — discovery. New customers who have never heard of your place find you there. This role is real and costly to replace.
The strategy that works is different: treat the app as an acquisition channel and WhatsApp as a repurchase channel. The customer arrives via the marketplace, has a good experience, and is invited — correctly and without violating platform rules — to order directly next time. The first order costs commission; the second, third, and tenth do not.
Continue reading for free
Fill out once and unlock this and all other blog articles. We don’t charge anything — we just want to know who we’re writing for.
Takes 20 seconds. Prefer to talk directly on WhatsApp?
The WhatsApp order flow
The mistake that breaks the operation is trying to take orders through free conversation during peak hours. The flow needs to be structured.
1. Always accessible menu. Fixed, updated link with price and photo — on the business profile, in the greeting message, and as a quick reply. An outdated menu image causes wrong orders and disputes at delivery.
2. Standardized order format. Always ask in the same order: items and quantities, full address with reference, payment method, if change is needed, and notes. An automatic message with this script drastically reduces back-and-forth.
3. Confirmation with price and delivery time. “Order confirmed: 2 large calabresa pizzas, R$ 118, delivery in about 45 minutes to Rua X, 220. Payment via Pix.” Explicit confirmation avoids 80% of problems at the door.
4. Payment before dispatch. Payment link or Pix with receipt reduces cancellations, eliminates change issues, and speeds up delivery. For recurring customers, payment on delivery can still be offered.
5. Dispatch notification. A message when the order leaves. It’s simple, costs nothing, and significantly reduces “where’s my order” calls, which are the biggest disruption during peak times.
6. Order recorded in customer history. This enables everything that follows — repurchase, preference, reactivation.
The bottleneck isn’t the order, it’s repetition. One attendant can process many standardized orders per hour; what stalls the operation is answering fifty times “what’s the delivery fee for neighborhood X.” Automate repeated information and keep the person for the order itself.
How not to turn chaos at peak times
Between 19h and 21h, everything happens at once. Four decisions save the operation.
Automate the three most repeated questions: delivery fee and time by neighborhood, payment methods, and menu. Alone, this usually solves half the message volume at peak.
Single queue with a defined owner. Nothing worse than two attendants answering the same order, or none. Each conversation needs a visible responsible party.
Response time as an absolute priority. In delivery, whoever takes five minutes to respond loses the order to whoever replies in thirty seconds — the customer’s decision is immediate and they’re hungry.
Honest pause when the kitchen is saturated. “We currently have a 60 minute queue” loses fewer customers than accepting and delaying. Complaints about delays are the biggest reputation destroyer in delivery.
The repurchase strategy
This is where WhatsApp beats any marketplace, because it allows you to talk to those who have already bought. Four moments, all with recorded permission.
Post-delivery, 40 minutes later: “Did everything arrive okay?” Short, sincere, no sales pitch. Resolves issues before they become bad reviews and opens space for praise.
Seventh day without ordering, for frequent customers: a light message with the week’s news. Frequency is the most valuable data you have — someone who ordered every Thursday and disappeared two weeks ago is a customer at risk, and no one notices that without records.
Reactivation in 30 days: for those who were customers and stopped. Here an offer makes sense, and the return rate is usually good because the memory of the experience still exists.
Dates and occasions: game day, holiday, heavy rain, city event. Contextual communication performs much better than generic promotions.
Two rules that preserve the channel: low frequency — at most two promotional communications per month — and easy opt-out, with the option to stop receiving respected immediately. Restaurants that send too much get blocked, and blocking lowers the number’s quality, as explained in the material on quality and restrictions.
How to bring customers from the app
Carefully, because platforms have rules about channel diversion — and the right way is to win, not capture.
Printed material in packaging. A card with the WhatsApp QR Code and a concrete reason to use it: a gift on the first direct order, faster delivery, channel-exclusive item.
Presence on social media and map profiles, with the chat button. Many customers search for the restaurant by name before opening the app.
Ads that open chat within a nearby radius, which usually have competitive cost per chat and bring the customer already inside your channel.
Simple loyalty program: every ten direct orders, one item free. Managed in the customer’s own history, no card or app needed.
The five numbers the restaurant should track
The advantage of owning your channel only materializes when someone looks at the data. Five indicators, reviewed weekly in fifteen minutes, support practically all direct delivery decisions.
Direct orders as a percentage of total. This metric measures the progress of the strategy. Moving from 8% to 30% over a year is a realistic goal and transforms operational results.
Average frequency per customer. How many orders each active customer makes per month. Increasing from 1.4 to 1.9 has a bigger impact on revenue than any acquisition campaign — and it’s exactly what the repurchase strategy produces.
Average ticket per channel. Direct orders usually have a higher ticket because the attendant naturally suggests sides and drinks, something the app doesn’t do. If your direct ticket is lower than the marketplace’s, there’s an opportunity for additional sales being missed in service.
Average first response time at peak. Measured between 19h and 21h, separate from the overall average. It’s the best indicator to predict lost orders.
Customers at risk. Those who ordered regularly and have gone two frequency periods without ordering. This is the most profitable list in the business and only exists when history is recorded.
None of these numbers appear in the marketplace report, and that’s exactly why they matter: owning your channel isn’t just about commission savings, it’s about the information it returns to the restaurant owner.
Mistakes that cause direct ordering to fail
Outdated menu. Wrong prices on WhatsApp cause embarrassment at delivery and customers who don’t return.
Slow response at peak. Delivery doesn’t forgive delays — the customer’s decision lasts minutes.
Not recording the order in history. Without this, the restaurant repeats the marketplace’s mistake: selling without knowing who bought.
Sending promotions every week. The fastest path to blocking and number quality decline.
Promising delivery times the kitchen can’t meet. Better to say 60 minutes and deliver in 45 than the opposite.
Direct ordering via WhatsApp doesn’t replace the marketplace: it transforms the customer the marketplace brought into a customer of the restaurant. It’s the difference between paying commission once and paying it forever.
Frequently Asked Questions
Is it worth taking orders via WhatsApp instead of the delivery app?
It’s valuable as a repurchase channel, not as a substitute. In a restaurant with R$ 90,000 monthly delivery and 22% commission, migrating 30% of volume to direct ordering generates between R$ 3,500 and R$ 5,500 in net monthly savings, after delivery, payment fees, and service time. The app remains important for discovering new customers — the strategy is to use it for acquisition and WhatsApp for subsequent purchases.
How to organize WhatsApp orders without chaos at peak times?
Standardize the flow: menu in a fixed, updated link; single order script with items, address with reference, payment method, and notes; explicit confirmation with price and delivery time; payment before dispatch whenever possible; and notification when the order is out for delivery. Automate the three most repeated questions — fee and time by neighborhood, payment methods, and menu — and maintain a single queue with a defined responsible person per conversation.
How to bring customers who order via the app to WhatsApp?
With invitation and reason, not capture. Include a card with QR Code and a concrete benefit for direct ordering in the packaging, keep the chat button active on map profiles and social media, use ads that open chat within a nearby radius, and offer a simple loyalty program managed in the customer’s own history. Remember that platforms have rules about channel diversion, so the way is to win the customer, not take them from the app.
Centralize service without changing your number
Connection via QR or official API, single queue with defined responsible, customer history stored by the company, and sales funnel on the same WhatsApp panel.
Continue reading
- Catalog and cart on WhatsApp
- Pix and payment link on WhatsApp
- Reactivation of inactive customers via WhatsApp
- WhatsApp ads that open chat: how not to waste budget
- WhatsApp service automation: what to automate and what to keep human
- Mass messaging on WhatsApp: how to do it without burning your number