For over a decade, food ordering apps have been how Britons get a takeaway—and now the weekly shop too. But the convenience costs: the apps take up to a third of each order in commission, and keep the customer. As platform fees now consume profit margins and budgets tighten, more UK retailers and restaurants are asking whether they can keep the orders without handing over the customer. Increasingly, the answer is yes. To unpack it, we spoke to Maxim Kozlov, lead product manager at Lightspeed Commerce, where he works on the company’s ecommerce site builder. His day job is building the tools merchants use to sell online, which gives him a close view of how businesses of every size are rethinking who controls the order.
Let’s start with the numbers, because they surprised me. Who’s actually winning UK food delivery?
It depends entirely on how you count. On the most recent breakdown from Lumina Intelligence—measuring delivery occasions—Uber Eats leads on 27.2%, with platforms run by the brands themselves right behind on 26.4%, Just Eat on 25.2% and Deliveroo on 16.2%. Trackers that measure order volume rather than occasions rank them differently, so treat any single figure as directional.
But the line that should make people sit up is the second one. Ordering direct—through a brand’s own app or a restaurant’s own website—is now neck-and-neck with the biggest app. That “fourth player” isn’t a single company at all. It’s everyone deciding to go direct, and it’s the part of the market that’s actually gaining ground. And this is happening in a market that’s still growing steadily, not shrinking—worth around £14.3 billion in 2025, up about 3% on the year, and forecast to climb past £15.8 billion by 2028. The pie keeps expanding.
Why would a business take on all that hassle when the platforms already have the customers?
Two reasons, and the first is just arithmetic. Commissions on the big platforms typically run between 25% and 35% of the order value. Delivery can be 30 to 50% of a restaurant’s revenue, but the platform takes 15 to 35% of every order—and for a lot of operators that’s the entire margin gone. You see the strain in the menu: the same dish that’s £12.95 ordered directly can be a couple of pounds more on the apps, because restaurants put their prices up to claw back what they pay in commission.
The second reason is the one people underrate, and it’s the one I care about most in my work: data. On the marketplace, the platform owns the relationship with the customer. You cook the meal, you never meet the diner, and you can’t market to them again. The appeal of the alternatives is that they put the business back in control of its own data and branding. The percentage is the headline; the customer relationship is the real prize.
Before we get to who’s winning, there was a whole category that lost badly—the rapid-grocery startups. What happened there?
That’s the cautionary tale that shapes everything else, so it’s worth dwelling on. At the 2021 peak, more than $18 billion of venture funding poured into speedy-grocery firms. Getir was the poster child—it hit a $12 billion valuation in 2022 and bought its rival Gorillas—and then in April 2024 it pulled out of the UK, Germany and the US within weeks. It joined what one writer nicely called a quick-commerce graveyard, alongside Gorillas, Flink, Weezy, Jiffy and Dija.
The model carried enormous costs, and once shoppers went back to physical stores, the demand simply wasn’t there to sustain it. The lesson the whole market absorbed: pure logistics, with no existing customer base and no reason for anyone to open the app, doesn’t survive on its own. The businesses thriving in delivery now are the ones that already had customers before a rider ever showed up.
What’s the structural shift that made “going direct” realistic for ordinary businesses?
The unbundling of the last mile. A marketplace really sells three things stitched together—the customers, the ordering technology, and the courier. What’s changed is that you can now buy each piece separately. And here’s the irony I find genuinely funny: the platforms themselves sell the piece that lets you escape them.
Look at how the supermarkets run their own rapid services—they lean on third-party courier networks like Stuart, Uber Express, Just Eat Go and Deliveroo Express for the last mile. Restaurants do the same. Uber Direct lets a business offer delivery from its own website for a flat fee per delivery—starting from £3—with no commission, keeping control of the customer. Even on the marketplace itself, you’re rewarded for taking back a piece: Just Eat costs roughly 14% if you use your own drivers, versus 28 to 30% if you use theirs. Once the hardest physical part—getting a hot bag across a city—becomes a service you can buy by the drop, the only question left is the one that really matters: who keeps the customer and their data?
So let’s map the responses. Start at the top—the big chains.
The deep-pocketed players increasingly keep everything. Tesco is the cleanest example. It refuses to list on the aggregator apps at all, so every rapid customer has to use its own service, Whoosh—which means Tesco keeps 100% of those customers at full margin with full data. It can hold that line because its brand and advertising bring in the customers on their own, which a small shop can’t match. And it’s working: Whoosh sales rose 47% year-on-year and Tesco holds roughly 37% of the UK online grocery market, inside a quick-commerce segment valued at about £2.4 billion in 2025 and forecast to grow around 10% a year to 2030. The threat that killed Getir became Tesco’s competitive edge.
And the supermarkets that don’t have Tesco’s scale?
They hedge—they play both channels. Sainsbury’s with Chop Chop, Morrisons with Now, the Co-op and others run their own platforms but also list on Deliveroo, keeping full margin and data on their own orders while giving up both on the aggregator orders. That’s the trade-off laid bare: your own channel is more profitable and tells you more about your customers, but the aggregator brings you people you’d never reach alone. Most are betting they can take some of each rather than choosing.
You mentioned the corner shop. That’s the tier people usually leave out. How does a small independent store get into this at all?
This is the part I find most exciting, because it’s where the technology genuinely levels things. An independent corner shop can’t drum up that kind of demand on its own—no one’s downloading its app. So specialist platforms have stepped into exactly that gap. Snappy Shopper brings customers to independents and symbol groups—stores under Nisa, SPAR, Premier, One Stop, Costcutter and Co-op—but, crucially, it gives control back in a way the big apps don’t. Stores keep command of their own pricing and range, get an in-house loyalty scheme, and around 80% of their delivery sales come from customers new to the store.
And delivery transforms the economics of a small shop, because the basket gets much bigger. The average order on the platform runs around £27, against a typical in-store spend nearer £10. They’ve even pushed into round-the-clock delivery for convenience stores. This is how local markets and independent retailers—the ones the delivery boom mostly forgot—are finally getting online: by paying a platform to bring them customers, one that’s built to hand the shopper back rather than keep them.
And restaurants, the original category?
They run the entire spectrum at once. At one end, independents that depend wholly on the marketplaces for visibility. At the other, operators who’ve gone fully direct on white-label logistics at a few pounds a drop. Most sit somewhere in between, taking back the pieces they can handle and paying for the rest. The point isn’t that one model wins—it’s that the choice is now far more granular.
This all sounds like bad news for the aggregators. Is it?
Not really. They’re doing two things at once. First, they’re expanding sideways into exactly these categories—Deliveroo and Uber Eats have both added groceries—and rebranding as local-commerce platforms rather than food couriers. Second, the 2025 consolidation matters here: the landscape is now dominated by Uber Eats, DoorDash, which owns Deliveroo, and Prosus, which owns Just Eat, and that gives them deeper pockets to defend the one thing that’s genuinely hard to copy. A corner shop can take an order and a courier firm can carry it, but neither can create the habit of customers reaching for an app in the first place. The platforms are leaning into white-label logistics, AI and subscription loyalty to protect that. Bringing in the customers was always the hard part to copy—not the riders.
So where does this all land in a few years?
Not direct or marketplace—a deliberate split that nearly everyone is converging on. Use the marketplace to get found and to win first-time customers, then move the regulars onto channels you control, where the margin and the data stay yours. How far you can push that depends on your brand: Tesco can refuse the apps outright, while the independent leans on Snappy Shopper for four-fifths of its delivery trade. But the logic is the same up and down the chain.
And quick commerce has found its real form—not as the cash-burning standalone businesses of 2021, but as one feature among many for retailers who already had customers. That’s the exact opposite of the original dream.
Last question—if you’re advising a small retailer or restaurant today, what’s the one thing you tell them?
Don’t treat it as a loyalty test between “the apps” and “going direct.” Treat it as a mix of channels. Use the marketplace for what it’s brilliant at—reach and getting you found—but build at least one channel you own, however modest, so that when someone orders from you a second time, you actually know who they are. The commission you pay is the visible cost. The customer you never get to keep is the expensive one.