Local Delivery Radar 2026: channel mix, tickets and margins of urban fulfillment

Finding 1 — How much does the aggregator really cost your margin?

The aggregator is the toll on traffic, not your margin partner: it moves volume in exchange for a commission that eats contribution per order.

That flow is real and it discovers your brand. But the mistake I see again and again is measuring gross sales instead of what survives the commission. A marketplace order with food cost at 35% leaves no EBITDA; with food cost inside the optimal range (National Restaurant Association) there is barely room for the fee. The aggregator gives discovery; your own site gives margin. They do not compete: they combine, and the owner sets the mix with cash numbers, not with hype.

Finding 2 — Unit economics breaks on the wrong metric

Delivery unit economics breaks when the owner measures revenue instead of contribution margin per order. The commission is a silent leak that only shows if you cost plate by plate. At 35% food cost, a commission order leaves no profit; between 28% and 32% —inside the optimal 28–35% band (National Restaurant Association)— the margin absorbs the aggregator's toll. Cash flow is the leading cause of financial stress and small-business closure (Inc.), and poorly costed delivery accelerates that bleed. At Masterestaurant we repeat it: the right metric is not how many orders come in, but how many contribution dollars remain after the fee. Delivery Hero moved €48.8 billion in GMV in 2024 (Delivery Hero, FY 2024); that volume only helps you if your cost sheet leaves room to pay the channel.

Finding 3 — Owned channel: less volume, more repeat business

The owned channel gives you margin and repeat business; the aggregator gives you volume and discovery. The winning strategy in 2026 does not pick one, it combines them. The repeat customer who orders through the marketplace is worth more once you migrate them to your own site, where you pay no commission. The meal delivery segment keeps growing, so the pie grows, but margin is fought for at the channel. The cash tactic: use the aggregator to capture, and include in every order an incentive —QR, coupon, magnet— that pushes the reorder through your own domain. Every second order you recover from the marketplace is full contribution, no toll.

Finding 4 — Dark kitchens and virtual brands: territory risk shifts

Virtual brands and dark kitchens shift territory risk: instead of an expensive storefront, you build a productive kitchen and multiply brands on the same line. The rise of dark kitchens in major cities is a sign the model is consolidating. Foodtech delivery and dark kitchens sit among the most funded verticals in the region (Bloomberg Línea). The margin trick lies in utilization: a kitchen producing two or three virtual brands spreads rent and utilities over more orders, and those fixed costs never load onto the plate —they go to break-even, not to food cost. The risk does not vanish, it moves: with no façade you capture no foot traffic, so you depend on the aggregator for discovery until you build your own brand and repeat base.

Finding 5 — Ticket and mix: raise the average before the volume

Raise the average ticket before chasing more orders: in delivery, a larger ticket dilutes the fixed cost of the run and the commission over a bigger base. The aggregator's fee is a percentage, but the logistics cost per order is nearly fixed, so a higher-value order yields more net contribution. Platforms like iFood connect a huge network of partner establishments across hundreds of cities in Brazil: in that sea of listings, the winner is not the cheapest but the one that builds combos and raises the average. Glovo's q-commerce already bills more than €1 billion a year, with grocery growing about 50% in 2024 (EU-Startups, 2025), proof that consumers widen the basket. Design the delivery menu with anchors, add-ons and high-margin drinks to push the ticket without giving away food.

Finding 6 — Automation and robots: where delivery cost is heading

Delivery cost is the next margin battle, and automation is already pushing it down. Autonomous delivery robots already complete several million deliveries a year in the US, and Serve Robotics has passed tens of thousands of commercial deliveries in Los Angeles. Zipline reached 1 million drone deliveries in April 2024, the first company to do so (Grand View Research). On ordering, White Castle's voice AI hits 90% order completion and roughly 60 seconds per order (SoundHound via Restaurant Dive, 2024), deployed across more than 100 drive-thrus (Restaurant Dive, 2024). A restaurant owner does not buy robots today, but reads the signal: whoever automates the last mile and order-taking lowers variable cost. Anchoring the business to the Masterestaurant framework means preparing the kitchen for that curve before commission crushes the margin.

Finding 7 — The 2026 delivery scorecard: how to read it

Read 2026 delivery as a three-column scorecard —volume, ticket and margin— and decide the mix with real public data, not impulse. Global GMV keeps growing: the major aggregators posted double-digit growth in 2024, and instant retail in Asia moved a comparable multi-billion-dollar volume. Demand exists; the question is how much margin you retain. This synthesis by Diego F. Parra and Masterestaurant orders that data by restaurant segment so the owner does not confuse market size with their own profitability. The cash rule is simple: the aggregator to discover, the owned channel to retain, the ticket to dilute fixed costs, and food cost under 32% as the line you do not cross. That is the mix that survives the silent commission.

Finding 8 — The differences that decide the margin

The aggregator gives you volume and discovery; the owned channel gives you margin and recurrence. They don't compete: they combine. The marketplace commission is the toll on traffic —Delivery Hero moved €48.8B in GMV in 2024 (Delivery Hero, FY 2024)— and it's only affordable if a food cost below 32% leaves room for it. Delivery unit economics break when the owner measures gross sales instead of contribution margin per order. A commission order with a 35% food cost leaves no EBITDA; with food cost inside the optimal range (National Restaurant Association) it does. The right metric is the margin that survives the silent commission. Virtual brands and dark kitchens change territory risk versus the storefront with a fixed address. Multiplying brands over one kitchen lowers fixed cost per brand, but saturates the marketplace and compresses the aggregators' AI recommendation shortlists.