Shared Physics
by Roman Kudryashov
Shared Physics

How Long Until Cloud Kitchens are the SEO-Hacking of Food Delivery?

Published on 5 min read

I've recently noticed more and more new restaurants popping up in my neighborhood.

I don't live in a food desert, but I have dietary restrictions that limit what and where I can eat (I eat vegetarian). I'm not alone in this: by some counts, approximately 15-20% of the population also have food sensitivities and dietary restrictions.

Since moving out of New York, the number of restaurants around me at any time that cater to dietary restrictions has gone way down, which means my care in selecting where to eat has needed to go way up.

That why I've been surprised to find so many new restaurants popping up with names that scream "order from me"! Things like Vegan Palace, Gooder Foods, El Vegeteriano, Urban Healthy, and so on (names changed for this post).

Some of these restaurants are real and I've eaten there. Some are real and I've driven by them. Others are not "real" in the sense that they are cloud kitchens/ghost kitchens. They don't except via delivery apps.

Cloud/ghost kitchens are an interesting business idea. There are two broad 'types' of these kitchens:

  • One type is when you have an industrial kitchen that is exclusively delivery-focused. There are no service-related costs – such as seating or wait staff – because the entire "restaurant" is only kitchen and staging space.
  • Another type is space cohabitation, where a second business is operated out of a primary business. Essentially, two restaurants share a space. An example is when you have a cafe in the morning and a bar in the evening, or two food stalls that might share a back-of-house kitchen team.

The overlap between the two is not insignificant. Industrial kitchen spaces can house multiple businesses, and restaurants are notoriously low-profit-margin businesses with expensive real estate footprints in high density areas that are optimal as delivery bases.

Which brings me back to dietary restrictions: when I search for food, I search for specific types of food. Lately I've noticed more restaurants that have an extremely limited menu (such as three or four different orderable items), supported by generic sides and addons (colas, fries). Or menus with many items that are really a variation on a theme (such as when a sandwich, a bowl, and a salad are all the same thing except for one changed ingredient - bread, grain, leaves).  

Some of these are so niche that they are hard to imagine as real businesses with retail footprints that could consistently support a volume of sales needed to pay for themselves (though who knows... smoothie shops apparently exist just fine).

But, I can easily imagine them as a cloud kitchen with low overheard and outsourced staff, not unlike a drop-shipping ecomm business.1  Or I can imagine them as a branded segmentation of a restaurant's larger menu, a selection of foods that would have otherwise been hidden or ignored.2

It makes me wonder the big question of this post: at which point will these cloud kitchens turn into a form of search-optimized listings spam, where "restaurants" are spun up around certain kinds of dietary searches and/or search terms, where a single restaurant segments their menu into five restaurants catering to different types of food, with each sub-brand carrying just a few menu items?

And then, what are the consequences of this?

I seek out dietary-exclusive restaurants because cross-contamination is a big deal, sometimes life-threateningly so. I know not to eat anything at certain fast food chains because even items such as fries and ice cream can include or be cooked in animal lard. If I go to an eatery that caters to a specific restriction, I have confidence that I am safe with any menu choice... a vegan place isn't likely to add fish oil or chicken stock to a sauce, or accidentally sweeping bits of bacon into your bowl from someone else's order. Grain free restaurants are carefully grain free.

But when I order from a cloud kitchen, all bets are off. Sure, there are food safety guidelines, but those cover the most egregious allergens – they emphasize the safety, which isn't useful when restrictions are due to digestive intolerance (rather than allergens), religious beliefs, or other reasons.

There are financial ramifications too. People talk about the the stock valuations of Beyond Meat and Impossible Foods and point at their adoption by fast food chains – first with excitement of broad market penetration, then with disappointment at low sales volumes. But what's the point of a vegan burger that's cooked in animal fat and topped with cheese? It's a curiosity at best, appealling to no one in particular. So without understanding the context of a kitchen, people with dietary restrictions tend to veer on the side of avoidance.

Cloud kitchens – in their broadest sense – have a lot of interesting potential. They're a low cost way to market-test the viability of a certain menu items. They can help surface foods by using sub brands to segment menus into standalone eatery units. They can be designed to appeal to otherwise underserved food-restricted groups. They can lower the total costs for a food production cycle and pass along those savings to customers (though who am I kidding... they'll probably pocket the profit).

But like any SEO-optimized landing page, they can also devolve into untrustworthy spam that clogs up search results when you're looking for something to eat. When does a cloud kitchen become just a landing page for a dietary preference search term? Sooner rather than later, I think.

Update, 2022.10.18

A week or so after writing this post, James Coleman in The Prepared highlighted MrBeast's "ghost franchises" innovation in restaurants. It's another interesting angle on the problem I'm describing:

MrBeast is a 24-year-old YouTube sensation who has amassed over 100 million subscribers. He is most well known for giving away large sums of money in challenges and stunts, but I got interested when he opened 300 fast food burger restaurants in a single day.

It turns out that they are actually “ghost franchises” that operate exclusively through delivery apps. In exchange for a cut of the sale (45%), MrBeast provides the branding, menu, and recipes to any commercial kitchen with excess capacity. Ghost franchises are now big business, powered by VC-backed startups like Virtual Dining Concepts. The kitchens don’t even need to retool. If you run a Mexican restaurant, for example, you can become a ghost franchise for Mario’s Tortas Lopez, a quick service concept from actor Mario Lopez. It seems like a win for small local restaurants, but are customers really “buying local” by supporting a national brand?


  1. What I mean is, all you need is a digital storefront and access to an industrial kitchen that can make menu items for you on demand. The delivery app will handle order intake and fulfillment, the kitchen will handle cook and prep, and so all you need to do is provide a recipe and a brand. It's like when ecommerce companies set up a website and then farm out production & fulfillment to random third party vendors that will just slap your logo on their product and call it a day.
  2. For example, if you have an expansive menu, there is no reason why the "sandwiches" section can't become its own "Burgers and Sandwiches" restaurant, the salads become their own joint, a hodgepodge of items that feature fish in them get arranged into a seafood cafe, and so on.

    Let's say you're a diner with a 10-page menu. In person, customers might visit because you have everything under the sun. But online, you're competing against everything else. A huge menu isn't an inherent plus when people are searching by type of food – so why not explode a single menu into five submenus and see which ones get traction, and which ones don't?

Thanks for reading

Was this useful? Interesting? Have something to add? Let me know.

Seriously, I love getting email and hearing from readers. Shoot me a note at and I promise I'll respond.

You can also sign up for email or RSS updates when there's a new post.