Article:

Ertugrul Elmas reckoned that a career designing fabrics would be a safe bet. Trends come and go but people always need trousers. He failed to appreciate that clothes could be made anywhere. When his employer moved its factories from Turkey to China in 2000, Elmas emigrated to London and found work in the textile industry. For five years, he earned a good living overseeing a cloth manufacturer in Romania. But after the country joined the European Union in 2007, many workers emigrated and the garment industry collapsed. Elmas decided to enter the restaurant business instead. After all, people would always need a good place to eat.

He opened Olives and Meze in 2014. Like Elmas, whose friends call him Eddie, the restaurant straddles two worlds, refashioning the Turkish grill for the bourgeoisie of Clapham, an upmarket suburb in south London. It was a success. With its sleek, steel-grey bar and open terrace, it quickly became the kind of place where you could take a date. Elmas didn’t think of it as somewhere that did takeaway – he assumed that was just for pizza, curry, and chow mein. But soon he noticed that most restaurants in the area were on one of three delivery apps: Deliveroo, Uber Eats, and Just Eat.

The restaurant was doing a good trade without offering delivery. But the graveyard of failed businesses is full of proprietors who refused to adapt. It turned out that Olives and Meze was exactly the type of place the delivery apps were keen to bring on board: it served restaurant-quality food that was more than a guilty pleasure. He could always stop if he wanted to, Elmas thought. Deliveroo took a commission of 25%, which seemed high. Once staffing, food costs, rent, and taxes were factored in, he was left with a margin of less than 10%. Still, he thought he could make it work.

Six years later, he was making a small but steady profit, enough to invest in a second restaurant, this time on a knobbly side street in Soho in central London. There would be a constant flow of passing tourists and shoppers and it was close enough to Theatreland to capture the pre-and post-show trade; he hoped that office workers, weary of yet another sandwich from Pret, would come for healthy meze and salads during the day. It was ready to open in March 2020. Then the pandemic struck. Suddenly Elmas had to pay rent on a central London site and the customers had vanished. He turned to Deliveroo once more, this time for sheer survival. The company was delighted to take him on. But the commission for the Soho branch would now be 35%.

Some restaurateurs describe delivery apps as an addiction. They know the habit will harm them, but it provides a short-term fix

As he recounts this chapter of his story, Elmas’s smile falters for a second. He talks with the patient, weary air of someone seeking absolution for his choices. He had taken reasonable business decisions throughout, yet the maths no longer added up. With every ominous ping of the tablet collating online orders, with every buzz of a receipt printing, with each helmeted courier waiting outside the restaurant, he was losing money.

His litany of complaints is familiar to any independent restaurant owner: rising rents, onerous business rates, and the proliferation of chains homogenising the high street. But Elmas felt particularly wounded by Deliveroo’s actions. “You need support from your business partner, and for us, Deliveroo is like a business partner,” he told me. “I feel they let us down when we really needed them.”

Uncertainty is inevitable when running a business, but to Elmas this seemed worse. In his short story, “The Lottery in Babylon”, Jorge Luis Borges describes how the titular lottery, unpopular at first, takes off only when punishments are added to the roster of prizes. The element of risk makes the lottery more attractive. As the stakes grow, the lottery’s scope expands to control the fates of citizens, until no one is certain which events it determines and which it does not. Food-delivery apps have been a lifeline to many restaurants throughout the pandemic, though some restaurateurs describe them as an addiction. They know the habit will harm them in the long term; nonetheless, it provides a short-term fix. But the relationship may be more accurately compared to the Babylonian lottery: though the apps chip away at their finances, restaurants feel compelled to buy yet more tickets.

Collin Wallace is one of the grandfathers of modern food delivery. As with many great inventions, his technological innovation was devised to solve a simple problem: how to order a sandwich to his classroom without his teacher noticing. As a 22-year-old at university in 2006, he created a simple computer program to turn a text sent from his mobile phone into a fax delivered to a restaurant. As he surreptitiously tapped out orders under his desk, class after class was disrupted by a stream of sandwich couriers. His lecturers ordered Wallace to take his ingenuity elsewhere.

Airports, stadiums, casinos and hotels all viewed his invention as a curiosity, not a business proposition. At the time there was little demand for a phone-operated food-delivery system. Websites such as Grubhub and Seamless hosted digitised restaurant menus – the online equivalent of a drawer stuffed with takeaway flyers. Everything changed in 2007, when Steve Jobs unveiled the iPhone and unleashed the potential of mobile apps. Suddenly Wallace’s technology was in demand.

Grubhub first courted Wallace’s company for the exclusive use of its technology, then bought it and appointed him head of innovation (“It’s just a title you give to the founder of a company you’ve just merged into your own company,” he told me). But Grubhub’s actions soon started to irk Wallace. “The goals became more monetary and more predatory,” he said. He felt a responsibility towards small immigrant-run restaurants, but instead, he saw them forced to swallow high commissions and hand over data about their customers. So he left.

To an outsider, the business models of food-delivery apps seem perplexing. In America, the market share of each of the Big Four – Grubhub, DoorDash, Uber Eats, and Postmates – has fluctuated over time, as those that made initial land grabs for city centres have been overtaken by other firms that focused on suburbia (a bet that the pandemic has handsomely rewarded). Grubhub used to dominate the market but in many places – apart from New York City – it has been superseded by DoorDash, which is now responsible for more than half of all food deliveries in America by value and is worth $70bn.

One thing has stayed constant: these companies all lose money. The apps know they aren’t going to make a profit yet. Their strategy is different: to steal restaurants’ customers. As Wallace wrote when commenting on a blog post last year, the apps “are simultaneously selling these same customers to your competitor across the street, but, don’t worry, they are also selling their customers to you.”

Delivery services have already attracted scrutiny – and often vilification – for using gig workers to fulfill orders, a model derived from Uber. They have also been criticised for adding restaurants without their permission or conspiring with Yelp, an online business directory, to push premium-rate numbers that charge restaurants a “referral fee” for each sale.

Such practices have often drawn attention away from something less overt and more profound: food-delivery apps are disrupting the restaurant industry itself. Restaurants have had to question where to base themselves, what to cook and, in a few cases, whether they will ever serve customers in person again. The results will have implications for what, where and when we eat in the future.

At their most basic, restaurants exist to fuel us. For most people, though, they’re about far more than that. Think of your favourite restaurant. Sure, you might picture a dish you’ve eaten. But it’s more likely that what actually comes to mind are all the experiences you’ve had there, whom you were with, how it felt to walk in through the door. For the city-dweller, a favourite restaurant might be your local or somewhere you travel to for special occasions. But it is invariably anchored to a particular place. A restaurant is a distillation of a city – when we enter a restaurant, we briefly enter another world, but it is one that reflects our surroundings.

His innovation was devised to solve a simple problem: how to order a sandwich to his classroom without his teacher noticing

Over the past decade, there has been a quiet attrition of the traditional notion of a restaurant. The romantic ideals of restaurant owners have played a role in this. To them, delivery has no place in the dining experience. At best, it’s a distraction. Delivery apps spotted this disdain and took away the burden, but in doing so they slowly changed customer behaviour. For food-delivery companies, a restaurant’s location is relevant only in determining the radius within which it can deliver. Having food couriered to your door used to be a replacement for home cooking. Increasingly, delivery apps are turning a night out into a night in. The coronavirus pandemic has intensified and accelerated this trend.

Before the apps, the economics of traditional takeaways were simple: restaurants charged the customer more than the cost of a driver. The first wave of food-delivery platforms served this market. In London companies such as Just Eat and the now-defunct Hungryhouse connected customers to food outlets: restaurants handled the logistics of delivery. This model was simple, boring and yielded a healthy profit. The downside was that it worked only for restaurants that already had their own drivers to ferry food.

The apps of the second wave, which include Uber Eats and Deliveroo, employ couriers as independent contractors to organise delivery. This has vastly expanded the range of meals you can order to your door. Suddenly the full diversity of a city’s cuisines – from Venezuelan to Iranian, from sushi to souvlaki – are a few taps of a screen away.

These apps charge restaurants a higher commission because they are doing more work, acting as an e-commerce portal, a marketing platform, and a courier. They are also changing the perception of restaurants. Establishments become increasingly indistinguishable as customers encounter them in the same way via the interface of an app. There is none of the communal energy of a dining room, the charm of a waiter, the opportunity for spontaneous generosity.

This matters, because these experiences lie at the core of restaurants’ historical business model: they foster loyalty and repeat custom. When ordering from home, you buy what you expect to eat. When you’re at a restaurant, you may find yourself wanting to prolong a meal, which means that you order more food and drinks, particularly alcohol, on which the greatest margins are to be found. A number of people have remarked to me that getting takeaway is significantly cheaper than eating out, even when delivery costs are factored in.

Delivery apps are not just a technological convenience, they are transforming dining culture. In Britain, Deliveroo best exemplifies this shift. “We think ‘food first’ not ‘logistics first’,” Will Shu, its founder, told me. “We view ourselves as a food company.” This attitude makes it stand out from its competitors: Uber Eats is a transportation company that realised it could make more money ferrying burgers than people; Just Eat has mainly stuck to predictable takeaway fare.

Shu is boyishly nerdy about restaurants. At a Vietnamese café on the Old Kent Road in south-east London, which I thought might intrigue him, he geeked out on fried intestines and regional styles of pho (all available on Deliveroo, naturally). He is an anomaly among bosses of delivery apps in that he genuinely loves food (“Eaten 60 chicken McNuggets in one sitting” was listed among his achievements on a Deliveroo marketing deck).

Like Wallace, Shu claims that his idea for a delivery app was born out of hunger. As an investment banker at Morgan Stanley in London, newly arrived from New York and pulling all-nighters at his desk, he says that he was dismayed when he realised that the capital’s takeaway options boiled down to a fridge-cold supermarket sandwich or a Burger King Whopper. Middling American cities had more choice. From the start Shu’s ambitions were clearly greater than a simple urge to obtain a better class of snack: he wanted to transform the eating habits of an entire city. Ten years on, his company is worth billions of dollars.

Uber Eats realised it could make more money ferrying burgers than people

Shu recognised that restaurants offer cultural capital as well as food. The likes of Uber Eats developed a partnership with McDonald’s; Shu focused instead on restaurants frequented by foodies. There are many reasons we choose to eat in a particular place – following a recommendation, to be part of the in-crowd or to feel virtuous by supporting a local business. Shu reckoned he could harness that loyalty. When delighted customers spotted a favourite destination on the app, some of that warm feeling was reflected onto Deliveroo. After using the app for a while, he hoped, they would think first of Deliveroo when they became hungry, rather than any restaurant in particular.

Deliveroo’s initial interest in independent outfits and bright, cutesy branding gave the company the appearance of a plucky upstart. But there’s only so long you can survive on cultural capital alone. In the past couple of years, each app has made a play from its own corner of the market to annex the center ground, aiming to encompass every type of restaurant. To do that requires a lot of money. In May 2019 Amazon stood ready to invest $575m to buy 16% of Deliveroo.

This was the next stage in becoming, in Shu’s words, “the definitive food company”. Amazon had disbanded its food venture, Amazon Restaurants, in 2018 and some reckoned Amazon wanted to use Deliveroo to regain a foothold within this potentially lucrative market. The investment would benefit Deliveroo, too. Amazon has unparalleled expertise in marketing, growing customer share, and most importantly, efficiently delivering things. Food delivery is often uneconomical because a single household puts in a single order from a single restaurant at any one time (once a meal has been ordered, an app often moves that restaurant to the top of the rankings for that neighborhood, in the hope of getting a second and third delivery). For the apps, the ability to cut delivery costs may ultimately account for their success or failure as a business.

Britain’s Competition and Markets Authority blocked the deal, citing concerns that it would damage future competition in both the restaurant and grocery-delivery markets. Amazon and Deliveroo continued to press their case, but by March 2020 the merger seemed dead.

The pandemic came to the rescue. After a sudden drop in food deliveries in the early weeks of lockdown, when restaurants shut entirely, Deliveroo declared that it might go bust. This allowed the company to present what is known as the “failing-firm defense”. It is rare to use the defense that bankruptcy would reduce competition more than a merger. But the competition authority, perhaps not wishing to kill a rare British tech success story, agreed to Amazon’s investment. The merger was approved on the basis that Amazon would be a minority shareholder, but the evidence presented to the authorities made it clear that Amazon wants to invest further in Deliveroo and perhaps acquire it outright, a strategy it has pursued in other sectors where it has failed to compete effectively.

“You don’t choose your hours. The weather and customers choose your hours”

While the competition authority was deliberating, more people than ever were stuck at home ordering food online. The apps kept them fed and watered. In the second quarter of 2020 Deliveroo made a profit for the first time, and it did so again in the third quarter. By the time the competition authority finally published its full ruling in August 2020, it had dropped all references to Deliveroo as a failing firm. Meanwhile, its couriers got a different kind of deliverance: previously they were treated, at best, as a nuisance; now, alongside doctors, nurses and bus drivers, they have been elevated into the pantheon of essential workers.

I applied to work part-time with Uber Eats in July 2020, to get a first-hand understanding of the gig economy (I tried signing up with Deliveroo too, but my application mysteriously remains pending, despite the company’s plans to double its workforce with 15,000 new couriers). Delivery apps boast that riders can set their own timetable. This is only half true, as one courier told me: “You can make money but only when it’s busy. You don’t choose your hours. The weather and customers choose your hours.” In the July heat, as people escaped their homes to enjoy the period between lockdowns, I understood what he meant. It was totally dead.

My first attempts at couriering were unsuccessful. The app would tell me “You are in a busy area” without supplying me with any jobs. Weekday lunchtimes took me to few restaurants, instead of sending me to Pret a Manger, McDonald’s, Burger King, McDonald’s, Subway, and McDonald’s again. So many McDonald’s, in fact, that I can recite by heart the notice the company sends each time you get a pick-up, informing riders that they’re forbidden from delivering burgers to schools.

Though the workings of the system can seem obscure, they’re not indecipherable. By sticking to the rules you can achieve some success at this game – and you’re certainly made to feel that this is a form of play. When you press “go” on the app, a virtual city spreads out before you, as though you’re launching a video game. As I cycled around, waiting to be summoned, I experienced a sudden flashback to playing Pokémon on my Gameboy as a child, zig-zagging interminably on the screen through the long grass, hoping to be pulled into a challenge.

You soon work out what constitutes strategic gameplay. The app bombards you with promotions and you’d do well to accept them. Ride your bike down to Croydon and pick up a 1.6x boost to your earnings. Activate “questing mode” at the weekend and you’ll get a bonus if you achieve your delivery target within a set window – completing ten trips in three hours, a number on the very edge of achievable, earns you an extra £20 ($27).

For the ruthless rider there are also in-game hacks, such as pedalling dangerously fast and running lights. I’ve never seen anyone cycle more recklessly than my fellow couriers. You can cancel collections if the staff tell you an order will take too long to prepare. Some couriers buy more tickets to their own lottery by running two apps simultaneously, which enables multiple pick-ups (though this is technically forbidden, it isn’t easy for the apps to prove you’ve breached the rules.) I even heard of riders who time their arrivals at hospitals to coincide with the end of a nursing shift, gambling that the NHS discounts some restaurants offer may lead to a windfall.

Food-delivery firms officially disapprove of these hacks but they’re the inevitable consequence of the incentives they offer. It may look like you’re gaming the system but ultimately you’re just greasing its wheels. As a courier, you’re focusing on your own earnings. The genius of the model is that the faster you deliver meals for Deliveroo and Uber Eats, the more money they’ll make too. You might think you’re beating the app. In reality, you’re competing with your fellow riders to nab the next delivery, though it can be hard to see them as rivals because you’re each stuck in your own virtual world.

Unlike traditional labor models, in which employees’ interests are roughly aligned, the gig economy pits workers against each other. Riders are generally too busy getting their jobs done to converse. There are few opportunities to build up the kind of solidarity that underpins collective bargaining. In this fragmented workplace, couriers are kept constantly busy under the panoptic gaze of the apps.

I was terrified that my epitaph would read, “He died trying to deliver a Paul’s Christmas Turkey Sandwich in a timely fashion”

Some people appreciate the gig-economy model, particularly the ease with which you can log on or off. Shu is a robust defender. He told me about his own experience working part-time as a courier while doing research before starting Deliveroo. The job paid £6 an hour cash-in-hand and he saw undocumented couriers sleeping on restaurant floors. In Shu’s eyes, Deliveroo legitimises a shadow economy previously populated by badly paid subaltern workers. Others argue that Deliveroo’s own policy of letting drivers lend their accounts to other riders perpetuates this shadow economy, since the lenders often charge their own commission.

Uber Eats and Deliveroo both claim to pay the minimum wage in Britain. In fact, they pay per delivery: each drop usually earns a rider £3-3.50. How much you earn depends on the time of day, the weather, and how skillfully you hack the system. During my time as a courier, I made £8.20 an hour on average including tips, a figure that falls beneath both the national living wage (£8.72 an hour) and London living wage (£10.75 an hour).

One icy November day, I headed towards the Isle of Dogs in London’s financial district to try out some hacks. This is hallowed ground for Deliveroo: it was conceived here in Morgan Stanley’s offices and its headquarters are now nearby. The area is dense with Chinese students and American bankers, two groups accustomed to food delivery. After a while, I learned a trick: if I accepted orders from Chinese names, I’d generally get a resident in the Isle of Dogs, meaning shorter delivery times.

It was still hard to earn a decent wage. Although the apps estimate how long it takes to reach a restaurant or customer, they don’t factor in finding the customer or picking up the food, which take up most of your time. The vast majority of deliveries I made as a courier were to high rises, so I had to navigate a maze of corridors, stairs and lifts. No one ever came down to collect their food. Part of me wonders if the popularity of delivery apps is symptomatic of the alienation from the high street engrained by the vertical distance between restaurant and customer. Part of me wonders if it’s just laziness.

Most troubling, I found my attitude towards the restaurants starting to shift. More than once I stood, fuming silently as staff fulfilled orders lethargically. I was often shunted to one side and made to wait even though I could see my package lingering. Sometimes I was banned from coming into a restaurant altogether. Customers usually greeted me with an apologetic thanks, tacitly acknowledging the indulgence of a meal taxied to their door, but restaurant staff could be curt to the point of rudeness.

I felt I was regarded as an inconvenience rather than the conduit between restaurant and customer. The apps have inserted themselves both between the restaurant and customer, and between the delivery driver and restaurant – in so doing, delivery firms have transformed complex social relationships into perfunctory transactions, leaving both restaurant and driver in a sea of mutual resentment.

Once, the app sent me pedaling onto a motorway. Realising this too late, I was terrified that my epitaph would read, “He died trying to deliver a turkey sandwich in a timely fashion”. Eventually, after weaving through back streets and getting lost in a post-industrial landscape of indistinguishable tower blocks, I reached the destined door and knocked. “Leave it outside,” came the shout. No face-to-face. No thanks. I took a picture of the food and logged off.

The Isle of Dogs may be the birthplace of food delivery in Britain, but Park Royal offers a glimpse of its future. This town-sized sprawl of manufacturing space has slowly expanded since the 1930s, sinuously stretching across three boroughs on the outskirts of west London. The Park Royal industrial estate is sometimes called “London’s kitchen”, at least by marketing companies: a third of all food prepared in the city is cooked here. The grey, windowless buildings conceal, among other delights, Syrian baklava bakers, trendy juice companies, and the McVitie’s biscuit factory. It is surprisingly quiet, the silence disturbed only by the hum of motorbikes and taxis.

You need to follow your nose to find the action. The comforting scent of Lebanese flatbreads, heady with za’atar, and the unmistakable smell of lamb fat burning on charcoal leads you to the Acton Business Centre. Its banal exterior hides a labyrinth of kitchens preparing all manner of cuisines: Venezuelan arepas stuffed with slow-cooked beef teased into loose strands; Trinidadian rotis folded around chickpeas; crayfish simmered in chillies, Sichuan peppercorns and garlic, their crevices laden with numbing, scarlet oil. This may sound like the latest trendy street-food market, but the customers remain snug at home, unaware that the multiplicity of restaurants offered by their apps is to be found in this single building.

Dark kitchens – simply defined as a kitchen with no public face – are not a new phenomenon. Caterers, wholesale bakeries, and market stalls have long used them. But dark kitchens that make food for delivery only are a recent development. Some people consider their very existence unpalatable. Tim Hayward, a food writer, wrote in the Financial Times that “restaurants that use dark kitchens obviously don’t want you to know that your delicious dinner was knocked out by semi-skilled kids on an industrial estate on the edge of town” (but is that any worse than a meal made by semi-skilled kids in a renowned restaurant or by semi-skilled kids paid a pittance by a local takeaway?).

“Kebabs suffer, dal tastes fine and biryani is delicious”

The new alliance of dark kitchens and delivery apps poses an existential threat to the restaurant, one of the last bastions of the British high street. Butchers, bakers and grocers have been replaced by supermarkets. Cinemas, cab companies and bank branches have shrunk into apps on a phone. The restaurant has always been a curious institution: a business masquerading as a cultural and social hub, or perhaps vice versa. Until now, it has been extraordinarily resilient.

The dark kitchen moves cooked food off the high street into liminal spaces such as Park Royal, where the culinary desires of a neighborhood can be gleaned from app data and efficiently satisfied by unseen kitchen hands. Until recently many owners were held back by a nostalgic attachment to the traditional restaurant. Even if you weren’t a Luddite, it was comforting to know that you could, theoretically, visit the place where your dinner was cooked. The pandemic severed that connection, and not just in London. Across the world people began ordering from dark kitchens: every restaurant was transformed into one.

“A year ago we would not be having this conversation,” Shamil Thakrar tells me, sounding vaguely embarrassed. “Our job is not to put food in your house but to bring you in.” Thakrar is a co-founder of Dishoom, a small, fashionable chain of restaurants in Britain known for its vivid evocation of Mumbai’s fading cafés. When the pandemic hit, Thakrar’s most pressing concern was “to keep every member of our staff employed”. Realising that his restaurants would need to deliver, he sent each dish on a journey round east London, shook it up a bit, and tasted it. He learned that “kebabs suffer, dal tastes fine and biryani is delicious.” He immediately halved his menu. Within three months of lockdown, Thakrar had opened four new branches, including one in Park Royal. These were all modular dark kitchens, a world away from its meticulously designed dining room.

Dishoom is a bellwether for which restaurants will thrive in delivery’s expanding empire. It is exactly the kind of upmarket restaurant that lends Deliveroo legitimacy and attracts customers used to dining out. Deliveroo works hard to court these chosen ones, offering them free dark-kitchen space in exchange for exclusivity. This helps apps counteract their biggest weakness: customers promiscuously flitting from platform to platform.

Shu thinks it’s too early to say whether the pandemic has irrevocably shifted customer behaviour, though he tells me an older demographic is now coming on board. Restaurateurs I spoke to suggest that coronavirus may hasten the demise of eat-in dining. Suddenly, restaurants are no longer in the business of hospitality, but of e-commerce.

For one indication of the lottery’s outcome, follow the bets. In July, Karma Kitchen, a British firm that rents out dark kitchens, attempted to raise £3m to fund its expansion. It ended up with £250m. The apps have been successful at attracting investment too. In America, DoorDash was valued at $15bn in June 2020 and is now worth at least triple that. Deliveroo’s much-anticipated flotation on the stock market in London is imminent.

These valuations have rocketed even though the apps make scant profit if any (a model that follows that of Silicon Valley darlings such as Uber and WeWork). Until last year DoorDash and Deliveroo were entirely loss-making. The companies and investors tolerated this, in the hope that they’d be the last app standing. Consolidation has already brought that moment closer. Last year Takeaway.com, a Dutch company, bought both Just Eat and Grubhub to form Just Eat Takeaway; Uber Eats swallowed up Postmates. In just under a year, a transatlantic seven has been whittled down to four.

The pandemic offered near-perfect conditions for the apps: people were stranded at home and restaurants had to turn to delivery. Deliveroo claims that in 2020 it made a profit in 11 of the 12 countries it operates in, though it hasn’t released precise figures. Yet even with all the stars aligned, profits were modest (it was still making huge losses in 2019).

Now that restaurants like Dishoom deliver, it has become clear that some establishments are more equal than others. Dishoom negotiates its own commission, which Thakrar admits is “very good”. Another well-known London restaurant group pays Deliveroo a commission of 19.5%, significantly lower than the 30-35% many others pay. For many restaurants, that margin is the difference between making a profit and breaking even.

“I think it’s unfair when some traders do the same amount of sales but have very different commissions,” says David Gutiérrez of Guasa, a Venezuelan street-food business also based in Park Royal. “I understand it if you’re doing much more volume. But I’m seeing similar businesses being charged vastly different commissions.” Guasa’s razor-thin margins forced him to create two additional virtual restaurants accessible solely through the apps – one selling tacos, the other Cuban food – operating out of the same kitchen space and using the same ingredients and suppliers to minimise his expenditure. Each is available on Uber Eats and Deliveroo, giving him six revenue streams.

As a restaurant, the easiest way to increase your chance of winning the lottery in Babylon is simply to buy more tickets. But is the prize worth winning? Gutierrez worries that his business may become unviable if his customers lie hidden behind the apps’ interface. Thakrar of Dishoom has a different concern: has he sacrificed the spirit of his restaurants in order to save them?

When the Laughing Heart, a wine bar and restaurant in Hackney in east London, shut its kitchen last March, its owner Charlie Mellor decided to eschew Deliveroo. The economics didn’t make sense and he was loth to place his business in hock to Amazon. “It’s just gross, man,” he told me. He tried out a new app that charged a lower commission, but immediately began to notice problems. Drivers would cancel if they were notified of a better job elsewhere and he fielded numerous complaints that food was lukewarm or late. Within ten days he received more criticism than in the previous three-and-a-half years.

The bearded and gregarious Mellor, formerly an opera singer, is known on London’s restaurant scene both for his hospitality and his refusal to compromise. He worried that the reputation he had cultivated over many years would be ruined in an instant by someone else’s shortfalls. So Mellor and his business partner Pavel Baskakov decided to build a basic website that would allow the Laughing Heart to process and deliver orders itself. Other restaurant owners took note. When enough of them expressed an interest in joining, he officially launched Big Night, a new delivery business that charges only 6.5% commission and commits to paying drivers the London living wage. Mellor hopes that Big Night will help stem that vicious spiral restaurants find themselves in as they compete over their most basic property: price.

By signing up to the main food-delivery apps, many restaurants have in effect had to absorb commission costs within their existing menu price, or even lower it for bundle deals offered by the app, wiping out any profit. Big Night is one of a number of platforms challenging the oligopoly of Deliveroo, Uber Eats and Just Eat in Britain. Since Big Night caters exclusively to independent restaurants, it will necessarily remain small. Other sectors of the restaurant industry are finding their own way to cultivate a digital relationship with customers. HungryPanda, for example, is used predominantly by Chinese students, who can order their native cuisine through a Mandarin interface. Such niches also help customers narrow the overwhelming number of options the biggest apps offer.

It would be unwise to romanticise the world before delivery as a paradise we will never return to

Shu, Deliveroo’s founder, acknowledges that something is lost when apps package up and anonymise thousands of restaurants. “There’s an emotional reason why you go to restaurants,” he told me. “Deliveroo needs to evolve from transactional to emotional.” He sounds vague when I ask what that means in practice. Thakrar, the owner of Dishoom, has already begun to consider manufacturing packaging that will ergonomically unfurl to draw the eater into the restaurant’s universe (though there’s a limit to how immersive this can get).

It would be unwise to romanticise the world before delivery as a lost paradise. Restaurants are partly responsible for letting the fox into the hen coop by outsourcing transportation of the food; part of the solution will be to reconcile delivery with their traditional business of preparing and serving food. The future of restaurant work may well be divided into back of house, front of house and out of house. The task ahead for apps such as Big Night is to make delivery drivers part of the hospitality service: they have become as fundamental to the running of a restaurant as a waiter or kitchen porter. Restaurant staff will need to treat them not as annoyances, or even stand-ins for customers, but as fellow employees.

Similarly, the rise of dark kitchens can be attributed to flaws in the old model. There is a logic to them in a city beset by high rents, where you can spend half a million pounds on a restaurant lease and refurbishment before a single customer has walked through the door. Some restaurant owners are thriving in this environment. One chef, who now runs his high-street restaurant as a delivery-only kitchen, says he has no plans to open to table service any time soon. Although the rhythm of the cooking is more awkward without the predictability of sittings, staff no longer have to pander to drunk or lingering customers. Business is good.

If restaurants are being transformed by delivery, it is also true that they have reconceived what we think of as deliverable. Some solutions simply aim at mitigation. Chips, as many restaurateurs have told me, steam and congeal as they travel, but demand for them has led to a trend for dirty fries, which marinate over the course of a long ride and make sogginess a virtue. A new vernacular of delivery food is emerging, imaginatively working within the constraints. On Big Night you can find rotisserie chicken from a Michelin-starred restaurant, an instant ramen kit and the Laughing Heart’s dumpling range, which can be finished at home in minutes.

Covid-era cuisine is not gourmet fare shoved into cartons, but a hybrid of restaurant-quality food, delivery-friendly dishes and home cooking. The dark-kitchen model encourages the rapid expansion of wannabe chains, yet it also allows the people who could never afford a high-street unit to open a restaurant. Amateur chefs and bakers can try out concepts without the risk of bankruptcy. From my room in south London, I can order hundreds of varieties of burger, but also artisanal panettone and Macedonian borek. One evening I cave and order some modern Trinidadian fare that is available only on Deliveroo. Within five minutes I get an automated text informing me that the order has been cancelled as a result of unforeseen volume. I can’t get through to the restaurant on the phone. After half an hour of failing to talk to a human, I give up. They’re busy, though who knows for how long.

Many restaurant owners are simply glad to cling on. Elmas, the proprietor of Olives and Meze, has gone all in on the lottery. During the first lockdown in the spring of 2020, his Soho branch joined Just Eat: it offered him a 25% commission, provided that he also granted it access to his Clapham restaurant. In the second lockdown in London in November he signed up to Uber Eats and a fourth app that turns out to be owned by Just Eat. None of this seems to have made much difference. When I spoke to him in December, he complained that he’d received only seven orders all week in Soho.

Elmas had seen the high street in Clapham hollow out even before coronavirus arrived but the pandemic accelerated the process. An Indian couple who owned a fried-chicken joint by the station are shutting up shop. A friend, an Egyptian restaurant owner, offered his advice: “Sell your restaurant and buy Amazon shares.”

Outside his restaurant, the street has been painted orange with Just Eat couriers on branded e-bikes, as they step up their war against Deliveroo and Uber Eats. I ask him which of the three companies he would bet on to prevail. “It doesn’t matter,” he says with a laugh, as though the question were absurd. “At the end of the day, we are all working to make a few people rich.”