In this episode of the McKinsey on Start-ups podcast, McKinsey executive editor Daniel Eisenberg speaks with McKinsey partners Vishwa Chandra and Victoria Lord about the recent rapid growth of the food delivery sector and what lies ahead for this complex ecosystem. An edited transcript of their conversation follows.
Daniel Eisenberg: Hello and welcome to McKinsey on Start-ups, I’m Daniel Eisenberg.
Many companies and industries have been upended by the pandemic over the past two years, but few consumer sectors have been more radically reshaped by the changes in our daily lives than food delivery. At the outset of COVID-19, amid lockdowns and physical-distancing requirements, delivery became a lifeline for the ailing restaurant industry. Two years later, food delivery has gone from a steadily growing but still small piece of the restaurant (and grocery) business to a veritable dynamo. The industry has doubled in the US, and is worth more than $150 billion globally.
Whereas restaurants largely used to handle the limited delivery options that existed, these days a complex ecosystem of players is involved, with an economic structure that is still evolving. The fact that the sector has mostly remained unprofitable hasn’t dimmed the appetites of investors who continue to plough funds into it; at the same time, global quick-delivery or q-commerce players are raising the stakes, promising the arrival of groceries, restaurant food or virtually anything else in only 10 or 15 minutes.
Today, we are excited to explore this dynamic sector with co-authors of a recent McKinsey article on its ongoing, rapid evolution, McKinsey partners Vishwa Chandra and Victoria Lord.1 In addition to their experienced consultants’ view of the food delivery business, both Vishwa and Victoria bring an operational, even personal, perspective to their work with the industry. Vishwa, who is based in the firm’s San Francisco office, was previously a member of the executive team at Instacart, one of the earliest grocery delivery platforms. Victoria, who is based in Miami, has served in commercial leadership roles for fast-growth startups in the food industry; she has a natural interest in the sector, as her family has owned a franchise restaurant for more than 50 years.
Vishwa, Victoria, it’s great to have you with us on the podcast today. Thanks so much for joining us.
Vishwa, what did the evolution of the food delivery ecosystem look like in the two to three years before the COVID-19 pandemic?
Vishwa Chandra: It’s a fascinating time to be in food. We are genuinely living through a once-in-a-lifetime evolution in how people eat and how people engage with food. And this has been playing out for a number of years.
Prior to the COVID-19 pandemic, the food sector was experiencing a good amount of growth in delivery. However, it seemed to be incremental progress. There had been some consolidation in delivery platforms, with new offerings being introduced, but it was steady growth in both restaurant delivery and grocery delivery.
There had also been some stabilization in the competition. The big four in the US—DoorDash, Grubhub, Postmates, and Uber Eats—were the primary delivery platforms. There was clear delineation between channels: restaurant delivery was restaurant delivery, grocery delivery was grocery delivery. We had seen some convenience delivery pop up. And while growth was continuing, in absolute terms, delivery wasn’t necessarily a major sales area.
Right before the pandemic, restaurant delivery probably accounted for $40 billion of a $600 billion industry. A large majority of that $40 billion was pizza delivery, which had been there for ages.
Victoria Lord: We were also starting to see experimentation with new business models around food and food delivery. For instance, dark kitchens were starting to emerge in the ecosystem, but they hadn’t fully landed. While there was some investor backing, restaurants were still hesitant to experiment in those ways. Those sorts of changes in business models, and that desire for experimentation, has been much more in the news in the past two years.
Daniel Eisenberg: So, you had a foundation, but you didn’t really have a fundamental shift until the pandemic forced lockdowns and restaurant closures. How have players in this delivery ecosystem responded to the pandemic? And how has it impacted the growth trajectory of the sector?
Vishwa Chandra: First and foremost, I think everyone in the sector was very conscious of ensuring safety for their employees, colleagues, and customers—and recognizing what everyone was going through.
After that, from a delivery perspective, there was almost five years of growth within five weeks. There was some volatility early on when people were a little apprehensive about ordering delivery. Once that quickly stabilized, there was a significant uptick. Suddenly the only means of reaching customers was through delivery. Restaurants had to adapt very quickly and evolve their offering, and make changes in how they thought about staffing, production, and the packaging their product—and their own economics. It required a fundamental rethink of their business and how they engage with their customers.
Daniel Eisenberg: As your article noted, the US market roughly doubled in size in the last 18 to 24 months.
Vishwa Chandra: We didn’t expect that to happen for at least another five years. The two things that were fascinating were the absolute amount of growth, and the speed at which that happened.
We talked a little bit about restaurants having to adapt their operations. But delivery platforms had to grow their delivery base as well. They had to recruit, onboard, and train a number of delivery drivers to meet demand. They had to make changes from a product perspective. Contactless delivery was never anticipated before. Suddenly, it was incredibly important.
To be honest, consumers had to adapt too. That meant changing their understanding of this new kind of customer experience, and almost redeveloping their relationship with their favorite brands.
There was also experimentation, and exposure to new brands and new cuisines, along with a sense of nostalgia. We saw a growth in demand for traditional cuisines that had always been part of delivery, but also for things that had resonance from people’s childhoods, from people’s experiences through life.
Daniel Eisenberg: In your recent article, you talk about three factors that will play a key role in the success of the various players— geographic competition, commission rates for restaurants, and driver delivery fees. Of those three, which do you expect to play the biggest role for the industry in the next two to three years?
Vishwa Chandra: It’s an interesting question. It depends who you ask.
From a delivery platform perspective, geographic competition is one of the primary drivers of their business. Just because three or four platforms are serving a market, it doesn’t mean that consumers are necessarily going to eat more food. They are sharing the same demand within an area. As consumers have started cross shopping, and shopping between platforms, the competition for their share of stomach and their share of wallet continues to grow.
The degree of competition and loyalty from customers in a particular market became incredibly important—and delivery platforms made deliberate efforts to attract that kind of following. You saw the growth of loyalty programs, and subscription-based models trying to lock consumers in, to ensure that a larger portion of their spend was on their platform.
For restaurants, it was all about commission rates. At the end of the day, those rates are what drive their economics. Whether someone eats in a restaurant or orders delivery, the food, labor, and packaging costs show only slight variations. But in a world where commission rates are the highest single expense item for a restaurant, that is the largest determinant for how sustainable this can be for them. Once you factor in everything from a delivery cost perspective, a tip perspective, and a surcharge perspective, that has an implication for the amount restaurants have to spend, and how much are they getting for what they are spending.
Daniel Eisenberg: I know sometimes I’ll order with one of the apps, but I’ll do pickup instead of delivery.
Vishwa Chandra: It’s interesting that you mention that. We talked about the necessity of evolving the product. There was a first wave of product evolution during the start of the pandemic and that was for things like contactless pickup.
As volume stabilized, and demand grew, the next wave of evolution involved product features. Pickup is one area where people are investing heavily because it is pushed from the consumers, and pushed from restaurants. As costs and fees are different, the experiences are different. Restaurants can maintain that connectivity with their customers—they can get people in and have a personal touch.
Daniel Eisenberg: There is also this growing trend of convergence in the types of products that delivery players are delivering, or want to deliver, whether it’s ready-to-eat or groceries. What is driving this trend? And what will it mean for the competitive landscape overall among these platforms?
Vishwa Chandra: We are definitely seeing a convergence. Traditional restaurant delivery customers and platforms are going into grocery, and grocery delivery platforms are getting into ready-to-eat and meal delivery, both directly and with their partners.
In the end it’s a battle for share of stomach. Prior to the last two years, consumers had thought about these channels very distinctly. Going out to eat, getting something delivered, and having ingredients to cook your own meal were very distinct occasions. As folks have spent more time at home in the last years, they have redefined their relationship with food as an experience. For many, it was the only experience that breaks up the day.
Some of this convergence is also driven by the question, is this a winner-takes-all market or not? We believe there will still be multi-marketplaces and multi-platforms. But there will be continued consolidation as fewer platforms gain consumer loyalty and are able to control a greater share of stomach.
Victoria Lord: Vishwa mentioned share of stomach, but as the delivery platforms start to expand into different categories, it’s no longer just share of stomach. It starts to become more share of wallet for a given consumer because now a consumer can log into their app and buy anything.
For instance, I can place an order for dinner tonight, for alcohol, groceries, and for sunscreen and pet food all through the same ecosystem. So, convergence means that these apps are making it easier for us to have more of our occasions and more of our shopping needs met through one ecosystem.
Think about what Amazon did in the retail space. It started with books and expanded into basically everything. Or what some of the super apps are starting to do in Latin America and in Asia. Those are some very interesting examples for how broad some of these app ecosystems can become over time.
Daniel Eisenberg: I know Getir is moving into California with 10-minute delivery and DoorDash is doing 15 minutes or less for grocery delivery in New York City. Is this going to be a major focus going forward? And is there a race to the bottom in terms of time that they’re going to promise?
Vishwa Chandra: The question becomes: In terms of absolute market size, how big is that? And that comes down to occasions. There is a convenience occasion. There is an emergency fill-in occasion where speed becomes important. But as you think about consumption patterns, and demand patterns, it is a smaller portion of the market. That’s why you see some platforms approaching it differently by asking, how do we actually serve up the right offering that meets the customers’ needs at that moment?
Whether you’re ordering restaurant delivery and want to add on a pint of ice cream or a six-pack of beer, or you’re suddenly out of baby products and diapers in the middle of the night, your ability to get that product at that moment of need with the right offering is what is going to be important.
Victoria Lord: The other thing I would add is consumer’s willingness to pay. You hit a limit at some point. I’m willing to pay a premium if I realize halfway through making a meal that I need something to finish it off. I’m less willing to pay a premium if I’m doing a bulk grocery order that I don’t necessarily need tonight, or even tomorrow, or the next day.
I think we’ll start hitting the boundaries on consumer’s willingness to pay across different categories and occasions, like Vishwa mentioned. That will be part of the limit as to how much is possible through these app ecosystems over time.
Daniel Eisenberg: Victoria, let’s shift our focus to the economic forces on some of the stakeholders. And in particular, let’s start with restaurants. How should restaurants think about balancing growth in delivery versus their core in-restaurant dining?
Victoria Lord: This is a tough question, and there isn’t one single answer. Restaurants really need to be deliberate. We’ve heard before, in other contexts, that not all revenue is good revenue—and that’s true in food delivery too.
If I were a restaurant owner, I’d be taking a close look at my economics for my restaurant, or my chain of restaurants. I’d ask, how much do I actually make on an average delivery order after all of my costs and the third-party fees? And that includes refunds, by the way. How does that compare to my in-store profit margin? And can I handle that delivery volume within my existing overhead structure and staffing level? Or am I starting to adversely affect my in-store business or incur additional operational cost to meet that delivery need? I think, especially now, many restaurants feel that they must participate in delivery. But the decision to do so is much more nuanced.
I can think of several restaurants in my local market that have chosen to use their own delivery fleet rather than participate with a third-party platform. They might require customers to order through their own website, even if they have a third party running last-mile logistics, or they offer pickup only. Those are all ways that restaurants can regain some control of the margin.
For the larger chains, I would be testing new business models. I would consider if, operationally, it makes sense to use a dark kitchen, or a commissary, that pulls the delivery out of the existing restaurants. Do smaller-format restaurants that are more streamlined for delivery and pickup make more sense? And I’d be working really hard on getting to good partnership terms with the third-party platforms, which have a real incentive to be able to feature great restaurants.
Daniel Eisenberg: You had an exhibit in the article that showed how the economics of a restaurant generating more and more of its revenues through delivery could be a double-edged sword. So, what is the impact you expect to see on restaurants’ profitability if more revenue is generated that way?
Victoria Lord: Restaurants traditionally make margins of seven to 22 percent. So, on a $35 order, you’re looking at $2.50 to $7.70 in margin. Delivery platform commissions are roughly 15 to 30 percent. I don’t even need to do the calculations for us to know that the math just simply doesn’t work.
While the revenue from delivery grows, profitability for the average restaurant declines. Of course, the more of the restaurant’s business that goes to delivery, the worse and more unsustainable that business model becomes. I have family members who are in the restaurant business, and they’ve decided against offering delivery from their restaurants for exactly this reason.
Daniel Eisenberg: Do you expect, at a certain point, that consumers will bear more of the cost of commissions? That they will pay more for delivery items? Or that restaurants will have different menu prices for delivery versus in-store dining?
Victoria Lord: There are two parts to that question. Restaurants can, and will, increase their prices online. Your average consumer is probably not pulling up the menu in the delivery app and comparing it with the one on the restaurant website—but I do because I’m curious. And I see anywhere from a few cents to a dollar difference in menu pricing depending on the menu item and the underlying menu cost. So, there is flexibility for restaurants to pass on some cost to consumers.
I think we’ve seen that in other spaces where you’re paying a premium for delivery, for convenience. It’s hard to say that consumers will bear all the cost. They’re going to bear some of it, but restaurants will continue to have to bear some of those costs because they cannot possibly pass on a 15 to 30 percent delivery platform commission on top of the 10 to 15 percent service charge, the delivery fee, and the tip that consumers are already paying.
We talk about the economics in the article, and we broke it down by each of the players, and the ecosystem. Consumers are already paying quite a bit. The restaurants are too, but I don’t think you’ll see a complete shift to the restaurants’ portion of those costs being passed on to the consumer.
Daniel Eisenberg: What are two or three key areas in which the platforms can drive down costs to achieve profitability going forward?
Victoria Lord: I hate to even talk about batching because, as a consumer, I don’t like thinking about my order getting picked up, with three stops on the way, and my food getting cold. But the math works for the platforms. The more orders that can be picked up at the same time and delivered in the same delivery run, the lower cost of delivery per order.
The economics there make sense. That is something that the platforms will continue to experiment with. And as their technology and the routing software get more sophisticated, they will be able to do this better and better over time.
And getting very tight on operational timing matters a ton for the platforms. All of them have quantified exactly how much money every minute costs. The easier they can make it for drivers to pick up orders without waiting, to drive tightly optimized routes, and to drop off orders, the better it is.
Vishwa mentioned the contactless delivery feature. It’s great for reducing the time it takes a driver to drop the order at the door. So, the delivery platforms benefit from that feature too.
Some of this is achievable through process enhancements, for instance through system and underlying technology enhancements on the delivery platform side, like the routing software I mentioned. And then some relies on restaurants doing operational and process improvements, like having dedicated pickup areas, or better signage.
Earlier we mentioned the cost of getting a customer, and the importance of loyalty and subscription programs. I think we will continue to see the platforms push on that, because when you lock me into your ecosystem as a consumer, and I prefer your app over others, it means I’m more likely to spend across those categories we talked about. It also means you don’t have to offer me additional promotions to get me into the app.
Daniel Eisenberg: How big a part of the market could subscription programs become? They not only drive loyalty, but also recurring revenue, which investors prize. Will that remain niche?
Vishwa Chandra: If you look at many delivery platforms, whether it’s on the restaurant delivery side or the grocery delivery side, subscriptions are already a large part of their offering. Not necessarily meal-kit subscriptions, but subscriptions where customers get preferences such as a reduced service charge, free delivery, or better access to promotions—these benefits have a lot of appeal for customers.
It’s a playbook that has been followed by many in the past, from a payment provider’s perspective, and from hospitality and airlines that have long-lasting loyalty programs. But I think it also becomes very interesting as consumers start saying: Who’s going to start paying for that?
At the end of the day, yes, the platform is rewarded with customer loyalty, but it does come at a cost. If your service fees are five percent versus 15 percent, that’s very meaningful. What we’re seeing is that it becomes a point of discussion between retailers, platforms and restaurants regarding the cost of that loyalty—where the benefit flows, and where the cost flows.
Daniel Eisenberg: Speaking of the consumer’s perspective, what shifts in price and experience can consumers expect from delivery services in the next few years?
Vishwa Chandra: I think you’ll see a couple of things. One is the continued channel blurring. Currently a restaurant platform looks like a restaurant platform, and a grocery platform looks like a grocery platform, with some add-ons. I think you’ll see that line blurring between each of them, with convergence into adjacent categories like alcohol, pharmacy, or office products, where you’re trying to capture a greater and greater share of a consumer’s daily consumption needs, whatever that consumption may be.
You’re also going to see a much more personalized, emotional connection that these platforms are going to try to make as they work to move away from just being a transactional platform. Whether that’s increased use of social, increased use of video, or a greater degree of engagement with each of the consumers. You already see that in other geographies where they’re ahead of what we see here in the US.
The last thing is a generational shift. Young families with kids drive the food industry, whether you’re a grocer or a restaurant chain. As more millennials and Gen Z’s start to have kids, and continue to progress in their professional lives, it’s going to be interesting to see how it plays out. We’ve already seen certain platforms making a bet on being more relevant to the next generation of customers.
Daniel Eisenberg: In the article you talk about untapped revenue pools. And you mention quite a few: Dark kitchens, virtual brands, and brand spin-offs. Can you talk briefly about which of these have shown the most promise to date?
Vishwa Chandra: Many of these are now at scale, whereas 12 or 18 months ago they were experiments. We are seeing quite a bit of growth in dark kitchens. This is a very operationally intensive business. Dark kitchens allow you to get some of the batching benefits, a greater chance of grouping orders together, than if you were a single kitchen or a single brand.
In addition to that, we’re also seeing more virtual brands. Retailers and restaurants are suddenly realizing that their brands have a resonance, and they can use them to expand. You have grocers launching ready-to-eat restaurant brands or full-service kitchens. Or, you’ve got existing restaurant brands realizing they can launch a third, fourth or fifth brand, leveraging much of the same equipment, ingredients and expertise that they have.
Daniel Eisenberg: Victoria, one of the other potentially promising opportunities you talk about is “menu engineering.” Can you expand on what that might look like, and what it will take to get there?
Victoria Lord: Menu engineering is fascinating. It’s very similar to what we see on a traditional restaurant’s menu board, where every single item has a role to play in terms of the level of its popularity, its profitability, whether it’s bringing folks into the restaurant, and whether its driving volume or driving margin.
But it’s much more dynamic because you’re able to use data to make well-informed decisions on a much more frequent basis. You can collect and use data on ordering patterns to revise your online menu regularly and promote those higher margin items up at the top, or promote items that you need to move quickly because you have significant supply of burgers sitting in the cooler. You can change your bundling strategy to drive volume, and you can change prices easily. You can also use that data to find the right assortment, the balance of number of items, that people are expecting to see.
You’re seeing this in the brick-and-mortar world too. Burger King, for example, just announced they’re moving to a streamlined, simplified menu to speed up some of their drive-thru operations. And they’re not the only chain that’s done this recently.
What this level of menu engineering will take, though, is deliberate data collection, deliberate data analysis, as well as the ability to make and implement quick decisions.
Some restaurants simply won’t have the scale and the capability to be able to do this. But some can. And the platforms could make this a service offering for restaurants because they have the data and they can compare across multiple restaurant banners, multiple markets, and multiple menus, and are able to provide rich insights.
Daniel Eisenberg: And they have the data scientists who can do it, right? Whereas the restaurants may not, except for maybe the big chains.
Victoria Lord: Exactly. I think the other part of this is personalization. We see a little bit of user-specific personalization in digital menus already. If you open the menu for your favorite restaurant, odds are good that you will see a section at the top that says, “items for you.” Those are chosen based on your past preferences and what the app has learned about the menu items that you like.
You could imagine a scenario where you would see an altogether different menu for the same restaurant if you pull up your app compared to what I would see.
You could potentially also see different pricing there at some point. This is much harder to implement operationally. I think the possibility of consumer blowback is much higher for something like this. So, who knows if that level of personalization and menu engineering would happen?
Daniel Eisenberg: We’ve talked already about the profitability challenges in this sector. But despite all of that, money continues to pour into the sector from investors. What is the outcome that investors are betting on, given that they’re conscious of all these economic challenges?
Vishwa Chandra: I think it comes down to the fact that food is still one of the largest sectors in the world. If we take the US, for example, there is a $2 trillion annual spend in aggregate between consumables, groceries, and restaurants, and not just on the delivery side. It is a very, very significant market that is going through a lot of change. Ultimately the platform that is able to influence the consumer is the one that will be able to drive significant value.
Daniel Eisenberg: With dark kitchens, the operations are happening behind the scenes, not in front of the customer. Is it safe to assume that a lot of the innovation in this sector is going to be on the back end?
Victoria Lord: I’ve been hearing much more about restaurants starting to trial their own dark kitchens where they are doing small, delivery-only storefronts, in some cases offering out part of that capacity as a third-party service. I think it will be very interesting to see, as that type of business model evolves, how much of it will continue to be with the third-party players who have existed to date versus the larger restaurant chains making their own investments in this space. The business model evolution behind this will also continue to be quite interesting not only on the platform side, but on the restaurant side as well.
Over the past couple of years, we have seen a tremendous amount of innovation at a much more accelerated pace than we expected. As a consumer, as someone who is passionate about this space, and who has family in the restaurant business, I am very excited to see what comes next.
Daniel Eisenberg: When you think about it, we’re all both participants and observers in this space, and it’s going to be fascinating to continue to watch and experience the evolution. This discussion has been great, and it’s made me hungry, as well. Vishwa and Victoria, thank you for joining us on the podcast today.
Vishwa Chandra: Our pleasure.
Victoria Lord: Our pleasure. Thanks, Daniel.
Daniel Eisenberg: Well, that’s it for today’s episode. Thanks again to our guests, McKinsey partners Vishwa Chandra and Victoria Lord.
As always, I also want to thank our McKinsey on Start-ups production team: Molly Karlan, Polly Noah, Sid Ramtri, Myron Shurgan, and Katie Znameroski.
And of course, thank you for listening. We hope you’ll join us again for McKinsey on Start-ups.