Join Us, or Die By Us

How the delivery platforms took the market, by an operator who shunned them and lost.

I refused delivery for years and I had good reasons ready for anyone who asked. That was the problem. They were reasons, not the reason. Most operators who manage to stay off will tell you it's commercial but for most of us it never was.

Depending on who I was talking to, I'd lead with the brand: 'I'm just not comfortable with what we do being diluted on Deliveroo'. Or the food: 'it goes against the grain when everything we do is about serving it as fresh as we can, and we want you to dine with us, not eat it lukewarm or soggy on the couch. Do call ahead and collect any time, though'. Well-versed lines, both of them, and they closed the conversation with a nod nearly every time. What they really were was an excuse for not looking hard at a revenue stream I hadn't bothered to consider with enough rigour. We were busy enough back then not to have to and the great migration of operators onto the platforms hadn't quite begun.

the platforms never won on food

The businesses the platforms decimated first weren't the ones fretting about brand. It was the take-out shops, and not just the curry houses and pizzerias that did both trades, but the delivery-only operations – set up purely to send food out or have it collected, with their own drivers, their own phone lines, their own regulars, and no commission due to anyone. Their food travelled because it had been built to travel and they weren't precious about any of it. Sadly, very few of them still exist in London that haven't now joined the platforms themselves.

And the platforms never went after their cooking – they went after their distribution, slipping it out from under them within a matter of months. They also dropped every competitor within three miles into the same scroll, so a customer who used to have a drawer full of takeaway menus now had a screen with forty options and no reason to be loyal to any of them. The day people stopped opening the drawer and started opening the app, everything those shops had spent years building was worth nothing. Knowing the streets, owning the drivers, being the number people called. Their only way back to a customer they'd earned over fifteen years now ran through the company that had just started charging 30% of gross sales for the privilege, and the drivers they'd built up were gig riders on the same app on inferior terms. It was a fairly straightforward proposition: join us or die by us – like a mob boss taking over a patch that was already yours and then charging you protection to keep working it.

That's the unflattering part the industry forgets or conveniently avoids. The platforms didn't win on food, service, or price. They won by becoming the place people look, and by the time any of us sat down to work out whether delivery suited the brand, the only question left was what commission we'd need to pay to be found.

how we got here

The platforms that marauded into the market are younger than most think, with the shift ultimately tipping on one thing: who controlled the riders.

Just Eat and Hungryhouse both turned up in 2006 and neither of them delivered a thing. They were order boards. You browsed the local takeaways, ordered online, and the restaurant sent it out on its own moped exactly as it always had. The platform took a cut for putting you in front of the customer, a fraction of what Deliveroo and Uber take now, and had nothing to do with the logistics. That left the operator intact and widened reach without touching distribution, which is exactly why it was never the model that was going to prevail.

Deliveroo arrived on the scene in 2013, initially with Will Shu himself running orders around Chelsea on a moped. Controlling the delivery itself instead of just the marketplace is the whole story of the last decade. Own the delivery and you own the customer, the data, the terms, the lot. Uber Eats followed with the same setup in 2016, making the order-board platforms, by then the norm, suddenly feel like an outdated convenience – and in a scramble to keep pace Just Eat spent a reported two hundred million pounds acquiring Hungryhouse in a deal announced in 2016 that took a full CMA phase 2 investigation before it cleared in late 2017. Consolidation at the bottom while the real shift had begun above.

Deliveroo Editions followed at around the same time. Off-site kitchens built into shipping containers stacked up across London industrial estates, letting a restaurant cook for postcodes it had no site in, with no front-of-house to pay for. Anyone paying attention could see it – the platform wasn't just a way to reach customers any more, it was where the food was being made.

Then the pandemic expedited the ascension further. With dining-in no longer an option, delivery boomed whilst a mass exodus of operators lined up to launch their home delivery menus onto one of the platforms. This was a significant turning point.

the hierarchy nobody writes down

Both operators and consumers know the pecking order, particularly in London, but you may not hear it said out loud. For the insatiable couch scroller looking for dinner, it's the fear, depending on the company, of coming across a snob. For operators, it's more likely that you signed with Deliveroo and might be running Uber Eats on the sly too.

Either way, Deliveroo is top. The best restaurants, sharper presentation and a significantly higher spend per head. Uber Eats is the broader church, far more of a mixed bag, good mainly for reach. Worth adding for the volume, but only if you're happy to give up the lower commission Deliveroo offers for exclusivity, which they police hard, scouring their partner pool for anyone in breach. Just Eat is the bottom, the order-board inheritance, the volume play. First one in, now cascaded down. It's a hierarchy of perception more than quality, but perception is the game when the customer is picking off a screen.

the reasons were all true

None of that ranking is why operators stay off though. The reasons they give are better than that, which is exactly what makes them so handy.

Who wants riders in the room? Helmets and hi-vis coming through the door mid-service, waving a ticket impatiently while your 7pm regulars are being sat. Who wants them out on the pavement queuing. Plenty of sites just can't run two channels through one pass without one of them suffering. Nowhere to hold the riders, nowhere to keep the packaging, nowhere to easily stage a bag for pick-up, and often all of it lands on the front side of the pass, in full view of the dining-in customer.

All true. All also impossible to disprove. You can hold any of those objections honestly, but you can also hold them because you don't want to be seen on the app, and from the outside there's no telling the two apart. There's no telling from the inside either, which is why you can sit inside the position quite happily for ten years. A reason that can never be wrong is a reason to be suspicious of yourself. I know which one I was – I just didn't know it then.

The food-won't-travel one is the only reason you can actually put to the test, so it's the one worth pushing on. Around three-quarters of the menu travels but the rest doesn't, so we leave it off the delivery menu. And some food genuinely can't be delivered with your head held high. Anything where the texture is the dish, anything that lives or dies in the thirty seconds after it leaves the pass, anything that carries on cooking in its own steam in transit. Some fried things arrive as a rumour of themselves with no amount of clever packaging able to save them. But a lot more survives than the refusers admit. Curries are a doddle, salads hold up far better than you'd think. Noodles are fine the moment you stop treating the soupy ones as the same dish and pack the broth separately.

This is where both sides get it wrong. The refusers say no to the lot on the strength of the quarter that doesn't make it. The desperate say yes to the lot, that quarter included, then can't work out why the reviews are so brutal. You aren't judged on your best delivery dish but on your worst, and on these apps the rating is the ranking and the ranking is whether anyone sees you at all in such a competitive market.

None of it is beneath anyone. Done properly it makes money even after commission, but properly is the word doing the work, and properly means attention to detail and consistency in taste and presentation as near as possible to the standard your dine-in customer expects. The packaging is your plating and the photo is your shopfront, and it all needs to reflect the quality you hold your concept to. Do it properly and delivery extends the brand. Do it badly and it's the most public way there is to cheapen it.

what actually breaks the position

Not an argument. Nobody reasons their way onto the platforms, or off them. What decides it is a bad year, and then another one.

Plenty of the operators now on the apps didn't change their minds – they ran out of the money that was paying for the opinion. I understood the feeling better than I'd like to admit. Waiting for a turn in the economy that never came, waiting for the room to fill back up the way it always had. At some point trying every conceivable option stops being strategy and turns into something more honest than that.

The catalyst was need. What made it bearable was watching operators I respected go and do it. Places I'd have sworn would never touch it turned up on Deliveroo and pulled it off, the brash unrefined platform branding somehow sitting oddly well next to the polish of the restaurant itself. When you're fielding a few sales approaches a week the first thing you do is check who else is on and which of your peers are in the mix, and the moment you see others ahead of you the objection quietly loses its grip. And it runs far higher than my peers. Nobu delivers black cod through Deliveroo from Mayfair. Roka, Dinings, Benares, Hakkasan, all on. Higher than that, even: as I write this, sitting in NW5, I can get a Claridge's Club sandwich to my door for forty pounds, or a bowl of the Connaught's granola with yoghurt and berries for twenty-six. Not a diffusion line, not a dark-kitchen sub-brand, but two Maybourne hotels at full room-service prices on the same app as everyone else. Somebody at the top of the group must have looked at its most untouchable names and decided the association it was meant to be protecting was worth less than the money.

And what I'd been protecting the business from turned out to be something I'd invented, more out of lack of necessity than principle, and it took me far too long to see it. Once I stopped defending the wall, it came down inside a fortnight, which is its own verdict on how well built it was.

what I actually gave in to

In hindsight I had no option. The food travelled, the snobbery was mine, and the numbers worked once you accepted the economics. Many operators running the industry-standard seventy per cent gross margin can expect a drop to nearer thirty on the same food once it goes out the door, after commission, packaging and paid platform promotion and discounting. The commission does most of that, and more than the headline number tells you, because it's charged on the gross, VAT included – so you're paying it on the fifth of the sale that was never yours. The VAT itself is a wash in any case though, as it lands the same on all your dining-in tables. Still a profit though, on covers that were far from guaranteed. So I signed. What I hadn't understood, and this is the actual confession, isn't that I joined too late – it's that I joined without really grasping what I was joining.

Not a partner. A layer that had put itself between the customer and the restaurant and had no intention of ever budging. And it kept the customer and it kept its cut of every single order. While I was feeling good about knocking a couple of points off my commission, the layer itself was being bought and sold over my head. Just Eat was sold to Prosus, a Dutch-listed investment house, and Deliveroo, the great British success story, was bought outright by US giant DoorDash in late 2025 for the best part of three billion pounds.

So huge swathes of the industry were forced onto the platforms, and that forced adoption was exactly what let them grow so dominant they became worth buying. And not a single pound of all that cooking stayed with the people who did the cooking – it went to the Americans and Dutch.

-