In a recent account of the restaurant industry, The Economist reported that Dishoom deliveries can include the same incense burned in its cafés, a QR code for a suggested playlist, and coriander and lime packed separately. The food is boxed; the atmosphere travels in fragments.
Dishoom’s chief executive, Brian Trollip, told the paper he hopes those details will entice customers to visit and experience the full restaurant for themselves.
That distinction matters. Dishoom is not pretending that a playlist and a stick of incense can reproduce an evening in one of its dining rooms. It is sending enough of the room home to make the missing part felt. The delivery becomes a trailer for the restaurant, leaving the customer to wonder whether they should now go and experience the full version for themselves.
It is a clever response to a problem now reaching far beyond Dishoom. Restaurants have become remarkably good at getting food to people who never enter them. Ordering has become easier, production faster and the journey from kitchen to customer more visible. But as the meal has learned to travel, the restaurant has begun to lose its hold over the occasion surrounding it. We can deliver the food; the harder question is what happens to the room, the brand and the culinary craft.
The restaurant became more productive when the diner stopped staying
One of the most consequential numbers in The Economist’s article appears some distance after Dishoom’s incense. After almost 30 years in which real labour productivity in American restaurants barely moved, it rose by 15% during the pandemic and remained at that higher level.
The obvious explanation is technology. Restaurants introduced smartphone ordering, delivery shelves, drive-through lanes and more automated production. Many learned to serve more customers with the same number of people. In a business facing punishing costs and narrow margins, that is a real achievement.
But the research beneath the number reaches a more revealing conclusion. Austan Goolsbee, Chad Syverson, Rebecca Goldgof and Joe Tatarka combined mobile-location and card-transaction data from more than 100,000 limited-service restaurants. They found that productivity rose most where customers spent less time. In the restaurants they studied, nearly the entire observed increase was associated with shorter visits, particularly those lasting less than ten minutes.
The restaurants had not simply become better at serving guests. More of their customers had stopped behaving like diners at all. They arrived briefly to collect an order, passed through a drive-through lane or never entered because the meal was carried away by somebody else.
This is genuine productivity. More meals or visits per worker represent a measurable operational gain. Yet the finding exposes a strange tension in the modern restaurant. If output is counted in transactions and labour hours, then time spent at the table appears to be something the system should reduce.
Going out to dinner becomes the inefficiency.
The dining room was never dead time
I started working in restaurants at 15, long before a phone could summon one to the door. The dining room was not scenery around the product. It was where much of the product was made. Before digital bookings, there was the paper reservation book: a week of scribbles, telephone numbers, party sizes, returning names and refusals. Some of the numbers were false, offered as insurance by customers who already knew they might not show up.
We still worried about turning covers. A table staying too long on a busy service could disrupt the reservations behind it and cost the restaurant money. There was never a period in which restaurant economics were suspended in favour of romance. But a table that lingered was not automatically a failure. It could also mean that people were comfortable, that another bottle had been ordered, that the conversation had found its own pace and that nobody was yet ready for the evening to end. That was a good sign for the restaurant, if not for the second sitting waiting for the same table, but it rarely required anything more dramatic than an extra glass of champagne for those patient enough to wait for the 10.30 booking.
The room produced things that were difficult to place on the bill: anticipation, recognition, atmosphere, accidental encounters and the particular electricity created when a collection of separate tables became a crowd. The meal mattered enormously, but so did the fact that it was happening there — and, of course, that everyone was talking about the food.
Takeaway, of course, is not a creation of the smartphone. Restaurants delivered food long before apps existed, just as fish and chips, pizza and Chinese food were carried home long before anyone used the phrase “off-premises”. Drive-through restaurants had already turned the customer’s time into an operational question.
The screen’s difference is one of scale, speed and intimacy. It places thousands of restaurants on one searchable shelf, remembers what we ordered, removes much of the conversation from the transaction and allows food to be compared, bought, paid for and tracked through a single interface. What was once an additional way of buying from a restaurant has become, for many customers, the restaurant’s front door.
According to the US National Restaurant Association, nearly 75% of restaurant traffic now takes place off-premises. That does not mean three quarters of American restaurant meals arrive by delivery rider; the figure also includes takeaway and drive-through orders. Its significance is more fundamental. Most encounters with a restaurant no longer require the customer to occupy one.
The customer changed before the kitchen
We tend to tell the story of restaurant technology from behind the counter. Software forecasts demand. Machines assemble meals. QR codes replace menus. Robots carry plates. Artificial intelligence promises to schedule staff, reduce waste and recognise the patterns hidden inside thousands of orders.
The productivity research suggests that the more important change began on the other side of the counter. The authors describe it as a change in the “technology” of demand. Customer behaviour altered the operating model before much of the machinery arrived to support it. The decisive machine was in the customer’s hand.
The screen compresses the distance between appetite and transaction. Hunger becomes a search, a row of images and a ranked set of possibilities. The order moves to the kitchen, payment disappears into the interface and a dot begins travelling across a map. At home, the lighting, music, company and entertainment are provided by entirely different systems. We can watch a food programme — or the influencers who prompted the order — on the same screen while we eat. The restaurant still cooks the dinner, but it no longer necessarily contains the occasion. In other words, it does not have to iron the napkins.
That convenience has real value. It feeds people who cannot or do not want to go out, extends the reach of restaurants and creates revenue beyond the number of seats in a room. The point is not that the delivered meal is somehow inauthentic. It is that the screen has changed the threshold against which the physical restaurant is judged: staying home has become frictionless, while going out has become a decision.
When eating out stops being casual
Trollip puts the problem plainly. “The casual dining element of the sector has disappeared because it’s no longer casual to decide to go out to eat any more,” he told The Economist.
Casual dining was built on a modest but culturally important proposition. It offered an ordinary meal outside the home without requiring a grand occasion. It occupied the useful space between fast food and ceremony: accessible enough for an unplanned Tuesday, distinctive enough to justify not cooking.
That decision now carries more weight. There is the journey, the booking, the weather, perhaps a taxi or a babysitter, followed by a bill inflated by the same food, labour, energy and property costs squeezing the restaurant. Home has improved as a competitor. Its food is more varied, its entertainment is endless and its interfaces have been refined to remove almost every moment in which someone might reconsider the purchase.
This is why the middle of the market feels particularly exposed. At one end, efficient formats compete on speed, value and reach. At the other, exceptional restaurants sell scarcity, theatre and a night that cannot be reproduced on the sofa. Casual dining carries the expense of maintaining a room while struggling to give customers a sufficiently strong reason to occupy it.
The economic pressure is severe. In the National Restaurant Association’s 2026 industry report, 42% of operators said their restaurant had not been profitable in 2025. Nobody running one of those businesses needs to be told that efficiency is unimportant. Technology that removes waste, improves forecasting or releases staff from repetitive administration may be essential.
The question is not whether restaurants should become more efficient. It is what they are becoming efficient at.
The room, sent in pieces
The Dishoom delivery bag brings us back to the point. The coriander and lime restore the last flash of freshness lost during the journey. The incense carries a recognisable scent into a different room. The playlist supplies the sound. The customer performs the final act of service by assembling those elements at home. Dishoom already describes dishes capable of travelling as “living room companions”.
It is hospitality translated into portable parts. Yet its purpose is not simply to make the home experience complete. If Trollip hopes it will lead people back to the café, the package is also admitting its own limit. It gives the customer a small sensory memory of the restaurant while preserving the value of the restaurant itself.
That is a more interesting response to delivery than trying to make the physical and digital experience identical. They are not identical. A meal eaten from a container while a series plays on television may be exactly what somebody wants, but it is not an inferior copy of dining out any more than dining out is an inefficient form of takeaway. They answer different needs, and the mistake is to imagine that the same measure of productivity can describe them both.
A kitchen designed to move food produces convenience, consistency and reach. A dining room designed to hold people can produce attention, company, memory and a small piece of public life. The first becomes more productive as the order moves through it quickly. The second may become more valuable because people want to remain.
In the programmable restaurant, the idea, menu and production system can increasingly be separated from the room. Dishoom reveals the same development from the customer’s side. Once dinner can reach us without the restaurant, the restaurant has to explain why we should still go to it.
The incense, playlist and garnish are a beautifully judged answer because they do not attempt to settle the question. They remind the customer that the meal came from somewhere with a smell, a sound, a history and people moving through it. They allow the food to travel while leaving something unresolved.
The screen can deliver dinner without the room. The future of the restaurant depends on making the room worth the journey.
Reporting and sources
This essay draws on reporting by The Economist in “The restaurant business is changing beyond recognition” (30 July 2026), including its account of Dishoom and comments from chief executive Brian Trollip. Restaurant productivity figures are also informed by research from Austan Goolsbee, Chad Syverson, Rebecca Goldgof and Joe Tatarka, published through the Becker Friedman Institute and the National Bureau of Economic Research. Off-premises dining data comes from the National Restaurant Association. References to Dishoom’s delivery offer are linked to the company’s own published material.
Links throughout the article lead to the original reporting and research wherever possible.
Illustrations are original artwork made for Eat Your Screen.