The biggest wave in consumer products right now has all the hallmarks of another bubble of misplaced investor expectations and sadly lower margins.
Cloud kitchens (the category, and not just CloudKitchens the startup service) is essentially WeWork for restaurant kitchens. Instead of buying an expensive restaurant site on a heavily walked street, a cloud kitchen is developed in a cheaper locale (an industrial district, perhaps), with dozens of kitchen stations that are individually rentable for short periods of time by chefs and restaurant proprietors.
It’s a market that has exploded this year. CloudKitchens, which has been funded by former Uber founder and CEO Travis Kalanick, is perhaps the most well-known example, but others are competing, and none more so than meal delivery companies. DoorDash announced that it was opening a shared kitchen in Redwood City just this week, Amazon has announced it is getting in the game and, around the world, companies like India-based transportation network Ola are building out their own shared kitchens.
That has led to laudatory headlines galore. Mike Isaac and David Yaffe-Bellany talk about “the rise of the virtual restaurant” at The New York Times, while Douglas Bell, contributing to Forbes, wrote that “Deliveroo’s Virtual Restaurant Model Will Eat The Food Service Industry.”
And there are not just headlines, but predictions of doom as well for millions of small-business restaurant owners. Mike Moritz, the famed partner at Sequoia, wrote in Financial Times earlier this year that:
The large chain restaurants that operate pick-up locations will be insulated from many of these services, as will the high-end restaurants that offer memorable experiences. But the local trattoria, taqueria, curry shop and sushi bar will be pressed to stay in business.
Latent in these pieces (there are dozens of them published on the web) lies a superficial storyline that’s appealing to the bright but not detail-oriented: that there are high software margins (or “cloud” margins, if you will) to come from a world in which kitchen space is suddenly shareable, and that’s going to lead to a complete disruption of restaurants as we know them.
It’s the same sort of storyline that propelled WeWork to meteoric heights before eventually crashing the last few weeks back down to reality. As Jesse Hempel wrote in Wired a few years ago about the shareable office startup: “Over time, this could be a much bigger opportunity than coworking spaces, one in which everything WeWork has built so far will simply feed an algorithm that will design a perfectly efficient approach to office space.”
Clearly, the AI algorithm for office efficiency (“WeWork Brain”?) wasn’t as profitable as hoped, with WeWork expected to lay off 500 software engineers in the coming weeks.
And yet despite the seeming collapse of WeWork and the destruction of its narrative, we still haven’t learned our lesson. As Isaac and Yaffe-Bellany discuss in their NYT piece, “No longer must restaurateurs rent space for a dining room. All they need is a kitchen — or even just part of one.” Now I know what the two mean here, but let’s be uncharitable for a moment: you can’t rent a part of a kitchen. No one rents the stovetop and not the prep area.
But it is that quickly slippery logic that can cause an entire industry to rise and eventually crumble. Just as with the whole “WeWork should really be valued as a software company” meme, the term “cloud kitchens” implies the flexibility (and I guess margins?) of data centers, when in reality, they couldn’t be further away in practice from them. Commercial kitchens require regulatory licenses and inspections, constant monitoring and maintenance, not to mention massive kitchen staffs (they aren’t automated kitchens!).
So let’s look at how margins and leverage play out for the different players. If you are the owner of one of these cloud kitchens, how exactly do you get any pricing leverage in the marketplace? Isaac and Yaffe-Bellany again write, “Diners who order from the apps may have no idea that the restaurant doesn’t physically exist.”
That sounds plausible, but if consumers don’t know where these restaurants physically are, what is stopping an owner from switching its kitchen to another “cloud”? In fact, why not just switch regularly and force a constant bidding war between different clouds? Unlike actual cloud infrastructure, where switching costs are often extremely prohibitive, the switching costs in kitchens seems rather minimal, perhaps as simple as packing up a box or two of ingredients and walking down the street.
That’s why we are seeing almost no innovation coming from early-stage startups in this space. Deliveroo, Uber Eats, DoorDash, Ola and more — let alone Amazon — are hardly under-funded startups.
In fact, this supposed rise of the cloud kitchen gets at the real crux of the matter: the true “expense” of restaurants isn’t rent or labor, but in fact is really marketing: how do you acquire and retain customers in one of the most competitive industries around?
Isaac and Yaffe-Bellany argue that restaurants will join these meal delivery platforms to market their foods. “…[T]hey can hang a shingle inside a meal-delivery app and market their food to the app’s customers, without the hassle and expense of hiring waiters or paying for furniture and tablecloths.”
Let me tell you from the world of media: Relying on other platforms to own your customers on your behalf and wait for “traffic” is a losing proposition, and one that I expect the vast majority of restaurant entrepreneurs to grok pretty quickly.
Instead, it’s the meal delivery companies themselves that will take advantage of this infrastructure, an admission that actually says something provocative about their business models: that they are essentially inter-changeable, and the only way to get margin leverage in the industry is to market and sell their own private-label brands.
For example, I get the same food delivered from the same restaurants regularly, but change the service based on which coupon is best this week (for me, that’s Uber Eats, which offered me $100 if I spent it by Friday). That inter-changeability makes it hard to build a durable, profitable business. Uber Eats, for instance, is expected to be unprofitable for another half decade or more, while Grubhub’s profit margins remain mired in the single digits.
The great hope for these companies is that cloud kitchens can fill the hole in the accounting math. Private brands drive large profits to grocery stores due to their higher margins, and the hope is that an Uber Burger or a DoorDash Pizza might do the same.
The question, of course, is whether consumers “just want food” or whether they specifically want the pad thai from that restaurant down the street they love because it is raining and they don’t want to walk to it. Food brands have a prodigiously long gestation period, since food choices are deeply personal and take time to shift. Just because these meal delivery platforms start offering a burger or a rice bowl doesn’t suddenly mean that consumers are going to flock to those options.
All of which takes us back to those misplaced investor expectations. Cloud kitchens is an interesting concept, and I have no doubt that we will see these sorts of business models for kitchens sprout up across urban cities as an option for some restaurant owners. I’m also sure that there will be at least one digital-only brand that becomes successful and is mentioned in every virtual restaurant article going forward as proof that this model is going to upend the restaurant industry.
But the reality is that none of the players here — not the cloud kitchen owners themselves, not the restaurant owners and not the meal delivery platforms — are going to transform their margin structures with this approach. Cloud kitchens is just adding more competition to one of the most competitive industries in the world, and that isn’t a path to leverage.
Brian Tarantina, second from left, joins the cast of “The Marvelous Mrs. Maisel” in accepting the award for outstanding performance by an ensemble in a comedy series at the Screen Actors Guild Awards in January 2019. (Photo by Matt Sayles/Invision/AP) 
Recent Comments