The Shift: Why Savvy Operators are Trading Marketplace

Restaurant delivery spent the past decade becoming wonderfully convenient for customers and alarmingly expensive for restaurants. One app could bring the diner, process the payment, send the order, dispatch the driver, and absorb up to 30% of the sale for its trouble.

That arrangement worked while delivery remained an extra channel. It looks different now. Nearly 75% of U.S. restaurant traffic takes place off-premises, according to the National Restaurant Association. Takeout, drive-thru, and delivery have moved from the margins of restaurant life to the center of it.

Operators across the U.S. and Canada have started responding with a subtler strategy than abandoning DoorDash or deleting Uber Eats. They are separating the functions that marketplaces once bundled together.

The marketplace can introduce the guest. A restaurant online ordering system can handle the next purchase. The POS can send the order into production. A local or national courier can complete the delivery.

The industry has spent years debating third-party delivery vs direct ordering as if restaurants must choose one side. The more consequential shift now involves using both, while deciding which channel should own each part of the transaction.

Restaurant delivery commissions now apply to a much larger share of the business, while restaurant margins remain painfully thin. DoorDash’s U.S. marketplace plans charge 15%, 25%, or 30% per delivery order. The difference between those tiers buys varying levels of reach and marketing support, but every percentage point still comes out of the restaurant’s subtotal.

On a $40 order, a 30% commission removes $12 before the restaurant pays for ingredients, labor, packaging, rent, insurance, software, or the occasional container of soup that arrives looking as if it has recently survived a maritime incident.

That tension reaches beyond the U.S. Statistics Canada reported a 4.1% operating profit margin for food services and drinking places in 2024. Operating expenses rose 4.8% to $99.6 billion, with the cost of goods sold accounting for 35.9% of total expenses.

The point is not that marketplaces charge for nothing. They supply consumer demand, payment infrastructure, support, merchandising, promotions, and access to large courier networks. A new restaurant can appear in front of thousands of hungry people without first building an audience from scratch.

The economics become shakier when the restaurant keeps paying an acquisition-level fee for a guest who has already ordered six times.

That repeat customer does not need another introduction. Yet the marketplace continues to sit between the diner and the restaurant, collecting a percentage while controlling the interface where the relationship lives.

This explains why searches for how to reduce DoorDash commissions for restaurants, how to reduce Uber Eats commissions, and how restaurants can avoid third-party delivery fees have become more urgent. Operators do not want delivery to disappear. They want the cost of delivery to reflect the value that each provider adds to each order.

For a first-time guest, discovery may justify the commission. For a regular who searches for the restaurant by name, the same fee can feel like paying a finder’s fee every Friday night.

Direct ordering gives restaurants a way to keep marketplace discovery while moving repeat demand into a channel they control. The restaurant no longer asks one platform to find the guest, process every future purchase, hold the customer record, and arrange every courier.

The marketplace still performs a useful role. It acts as a crowded digital food court where customers compare cuisines, prices, ratings, delivery times, and promotions. Restaurants would struggle to recreate that audience individually.

Direct online ordering for restaurants serves a different moment. The guest already knows the brand. They remember the birria, the pad thai, or the unusually persuasive chocolate cake. They need a quick route back.

That route might begin with a branded website, an app, a loyalty message, a receipt link, or a QR code on the bag. The direct channel removes competitor listings from the screen and lets the restaurant shape the menu, merchandising, offers, and post-order communication.

Paytronix reported in its 2024 Online Ordering Report that guests using first-party channels ordered 35% more items per check than third-party guests. Its research also links first-party ordering with higher purchase frequency and stronger loyalty participation.

Those findings help explain the appeal of first-party restaurant ordering. The value extends beyond a lower fee. A direct order can connect customer identity, location, purchase history, loyalty activity, and marketing consent inside the restaurant’s own technology stack.

That does not grant the restaurant unlimited rights over customer information. Privacy rules and vendor terms still apply. It does give the operator more control over the restaurant customer relationship than a ticket carrying a first name, a list of modifiers, and very little else.

The language around commission-free online ordering can create confusion here. DoorDash offers commission-free Online Ordering for orders placed through restaurant-owned channels, while Uber describes Webshop as a customizable commission-free ordering site. Both models may still involve payment-processing, software, marketing, or fulfillment costs.

Commission-free ordering means the restaurant avoids the marketplace percentage on that direct transaction. It does not mean card networks, software providers, and couriers have all decided to volunteer.

The deeper shift concerns ownership of the journey. Branded online ordering keeps the restaurant visible from menu browsing through confirmation. The guest remembers where they ordered, not only which app delivered it.

That difference matters when operators want to grow direct restaurant orders or convert marketplace customers to direct orders. The direct experience must feel at least as easy as the marketplace. Loyalty points cannot rescue a slow website, a broken modifier menu, or a checkout page that requests the emotional commitment of a mortgage application.

White-label delivery lets restaurants accept orders through their own channels and hire an external courier only for fulfillment. The customer sees the restaurant’s brand, while a delivery service works behind the scenes through an API or connected dispatch platform.

This model answers a question operators once considered impractical: how to offer restaurant delivery without third-party apps and without hiring, scheduling, insuring, and managing a private fleet.

Uber Direct, for example, handles orders that originate on a business’s website, app, or phone rather than inside the Uber Eats marketplace. DoorDash Drive lets connected applications request a quote, confirm coverage, receive a delivery price, and estimate pickup and drop-off times before creating the job.

The distinction between white-label delivery vs third-party delivery sits at the front door. In a marketplace order, the delivery platform owns the storefront. In white-label delivery for restaurants, the restaurant owns the storefront and buys the last mile as a separate service.

This produces a different fee structure. Marketplace commissions usually rise with the check size. White-label fulfillment often uses a per-delivery or distance-based charge. DoorDash Drive, for example, calculates fees according to the distance between pickup and drop-off, with a published base rate of $9.75 for deliveries within five miles.

That does not make every direct order cheaper. A small ticket traveling several miles can produce weak economics through either model. A large family order within a compact delivery zone creates a more favorable comparison because a percentage commission rises with the sale while the courier fee follows another formula.

Searches for flat-fee restaurant delivery and commission-free delivery options for restaurants often collapse ordering and fulfillment into one number. Operators need to separate them. Payment processing covers the transaction. The ordering platform manages the storefront. The courier fee pays for transportation. The restaurant may also fund discounts, loyalty rewards, or part of the customer’s delivery charge.

Once those costs become visible, the restaurant can make a more deliberate choice.

A downtown location might use a regional bike courier for short trips and a national on-demand network for wider zones. A multi-location group might assign providers according to coverage, fee, promised pickup time, or order value. A kitchen might use its own drivers during the lunch rush and delivery-as-a-service for restaurants after 8 p.m.

This approach turns direct restaurant delivery into a flexible supply chain rather than a second career in fleet management.

The POS increasingly coordinates orders that begin far beyond the counter. A modern restaurant POS with online ordering must connect menus, modifiers, prices, kitchen routing, payment records, store availability, courier status, and customer updates before dinner leaves the building.

That makes POS-integrated delivery less visible to the guest and more consequential to the operator.

When a direct order arrives, the system must recognize the restaurant and location, translate the menu data, send the correct items to the correct production stations, calculate prep time, request delivery, and return status updates. The guest sees a confirmation screen. Behind it, several systems conduct a small diplomatic summit.

Automated courier dispatch adds another decision. Instead of sending every order to one provider, restaurant delivery management software can compare eligible services according to coverage, price, timing, and restaurant rules.

Olo says its Dispatch network includes more than 27 delivery service providers, while 96% of its customers’ restaurants can access at least two providers. That degree of coverage makes multi-carrier delivery management possible without asking restaurant staff to open several dispatch portals during service.

Automatic courier assignment can favor the lowest eligible price, but cost alone rarely tells the full story. A provider that saves $1 and adds 18 minutes to the trip may not look so economical after the fries reach the guest.

Useful last-mile delivery orchestration also considers pickup timing, travel distance, service area, courier availability, cuisine type, and fallback coverage. The system can send a short-distance order to a local courier integration while reserving a national network for a wider delivery zone.

This connective layer has become its own category of restaurant delivery software. It sits between the restaurant’s POS, direct ordering channels, marketplaces, and delivery providers, translating data that each system describes differently.

KitchenHub operates in this middle layer. Rather than replacing the POS or becoming another consumer marketplace, it standardizes orders and menus across providers, maps items and modifiers, supports virtual stores, and passes delivery information between restaurant systems and courier services. Its integration framework includes national providers such as DoorDash Drive alongside services such as Relay and DeliverLogic.

The role feels closer to plumbing than theater, which is usually a compliment in restaurant technology. The best delivery service provider integration does not ask a cashier to admire it. It gets the correct order into the kitchen and the correct courier to the door.

For POS resellers, last-mile companies, virtual-brand platforms, and local delivery providers, that middle layer also removes a development burden. A standardized integration can connect several ordering and fulfillment services without forcing each partner to maintain a separate data model for every marketplace, menu, webhook, and order status.

The restaurant receives a more coherent workflow. The technology provider gets room to expand its offering. The kitchen gets fewer reasons to keep six tablets lined up like a small, blinking jury.

Restaurants are not staging a mass escape from DoorDash and Uber Eats. They are reducing dependency on delivery apps by giving marketplaces a narrower job: introduce new demand, serve convenience-driven customers, and extend geographic reach.

Direct channels can then carry more repeat business. White-label fulfillment can handle the courier. The POS and integration layer can keep the kitchen, menu, order, and driver working from the same version of reality.

The platforms themselves now support this hybrid structure. Uber distinguishes between its consumer marketplace and Uber Direct for orders that originate through merchant-owned channels. DoorDash offers Marketplace, Online Ordering, and Drive as separate products.

The market has quietly moved beyond a simple contest between direct ordering vs delivery apps for restaurants. Operators can now use each service where its economics make sense.

A new customer searching “Thai food near me” may arrive through a marketplace. The restaurant pays for discovery and reaches someone it could not reach alone.

A regular customer who knows the restaurant by name can use a first-party channel. The restaurant can retain more revenue from delivery orders, connect the purchase to loyalty, and communicate with the guest after the meal.

A courier network can complete either delivery. The driver does not need to own the order’s storefront.

This modular model also gives local delivery companies a better chance to participate. Through delivery service provider integration, a regional fleet can receive structured jobs from restaurant systems rather than asking every location to adopt another standalone dashboard.

For operators, the goal is not to keep 100% of every sale. Restaurants will continue paying for technology, payment processing, logistics, marketing, and customer acquisition. The goal is to understand which service earned which fee.

That distinction may sound almost quaint in an industry trained to accept bundled platforms as permanent facts of life. Yet it changes who controls the restaurant customer relationship, who can own customer data, and who decides where the next delivery order goes.

The future of restaurant delivery may still include the same familiar apps; they will simply occupy less of the picture.

The marketplace finds the guest. The restaurant earns the return visit. The software chooses the route.