👻 Five Whys—Ghost Kitchens

Sep 09, 2019

(1) Why should I care?

If you’ve ever ordered delivery from an app—like Uber Eats, Foodora, or Grubhub—chances are you’ve seen a ghost. You may even have eaten their food!

“Ghost kitchens” are a recent trend in the restaurant industry, a new style of restaurants emerging in response to growth in pickup and delivery demand.

Some people use the terms “dark kitchens” or “cloud kitchens”. For our purposes:

“Ghost kitchens” are restaurant-like foodservice operations that produce food for one or more restaurant brands not advertised at the premises
“Cloud kitchens” are coworking-like real estate operations that sell shared or dedicated access to commercial-grade kitchens

By keeping these concepts separate, we can talk about how this building in Los Angeles (operated by Travis Kalanick-owned Cloud Kitchens [1]) is a cloud kitchen, out of which several ghost kitchens (like Sweetgreen) operate:

(Look at all those vents!)

(2) Why should I listen to you?

I founded a company that helps foodservice operations manage, model, and optimize their recipes and menus.

Since ghost kitchens grow profits by developing, testing, and distributing new menu concepts—we’ll explore how that works in the next section—they’re not just an interesting new trend, they’re important to my work.

We even discovered how to catch a ghost (kitchen!), and can teach you too; let’s start with this hypothesis:

GIVEN: per our definition, ghost kitchens produce food for one or more restaurant brands; and
ASSUMING: ghost kitchens can grow profits by developing, testing, and distributing new menu concepts;
THEN: many ghost kitchens will have two or more menus registered to the same address on a delivery app.

While it won’t validate the hypothesis, we could test its viability by looking for delivery apps offering different menus at the same address [2]:

(All three pizza concepts belong to Family Style [3], a portfolio of brands “Built For Delivery”)

This heuristic has helped us identify hundreds of ghost kitchens in major markets. We compiled a dataset using popular delivery apps, and have used it to analyze cities like Los Angeles, San Francisco, New York, Chicago, Boston, Austin, Toronto, and more.

So as not to risk a swarm of copycat crawlers, I will not describe the methodology we used to compile the dataset any further. If you would like to see more of what we’ve found though, please do get in touch 🤓

(3) Why do ghost kitchens promote multiple menus?

To answer this we need to explore ghost kitchen economics. Let’s examine how running a ghost kitchen might reduce costs and increase revenue compared to a traditional restaurant.

COST: Running a ghost kitchen affects the restaurant’s COGS (cost of goods sold) via labor costs and food costs. The model also affects SG&A (sales, general, and administrative expenses) through rent and other fixed costs.

Labor costs—Don’t be deceived by a lack of FOH (front-of-house) staff; delivery app commissions (~20% - 30% of gross receipts) match common labor cost targets [4]. Ghost kitchens instead reduce labor costs by preparing larger batches of base ingredients, to be reused across multiple dishes and menu concepts.
Food costs—Just like labor costs, ghost kitchens can reduce food costs by designing their menu(s) around ingredient reuse and negotiating economies of scale with their vendors. Additionally, operators can turn menu concepts on and off as they become more / less economical due to e.g. seasonality.

“If I just decided to close the restaurant one day, and you drove over and it was closed, you’d be pissed. But if you’re looking for [one of my restaurants] in Uber Eats and you can’t find it because I turned it off, well, you’re not pissed. You just order something else.” —Eric Greenspan [5]

Fixed costs—ghost kitchens are excellent at amortizing fixed costs. For one thing, 100% of the rent pays for food production needs (i.e. no FOH). And again, with several menu concepts (like fish & chips, wings, and pizza) you can serve different clientele from the same location. Last, in areas where late-night foot traffic may not support a 24h “restaurant”, you might still be able to amortize fixed costs around the clock—just like this creative operator:

“One day, by accident, we left our tablet menu open for orders online for 24 hours, while we were closed, and when I came in the next morning, we had 40 missed orders. That’s when a light went on in my head, that delivery is a big thing.” —George Kottas [6]

REVENUE: Ghost kitchens depend on demand partners (like delivery apps) to sell their food. How does this affect typical strategies to grow revenue, like adding locations; increasing prices; and increasing traffic?

Adding locations—if a menu is profitable somewhere, it can probably be profitable somewhere else. Ghost kitchens can further reduce cost and risk by renting space from a cloud kitchen, before deciding if it’s worth adding a dedicated home in the new market.
Increasing prices—it is trivial for a ghost kitchen to test how menu pricing affects demand, by either changing prices for existing menus or creating a “new” restaurant with different prices.

Increasing traffic—delivery marketplaces can be competitive; one address we searched in Los Angeles showed 1300 restaurants available for delivery! Ghost kitchens compete by covering more cuisines (e.g. pizza, wings, souvlaki, fish & chips) and/or more of a cuisine (e.g. eight “different” pizza restaurants). For example, this NYC ghost kitchen markets 10 separate menus on delivery apps:

(None of the 10 are “Juni’s”!)

If you would like to explore further, I suggest starting from Cloud Kitchens’ “Delivery-Only” page [7]; I think their segmentation and sales copy is excellent.

If you are interested in getting your restaurant listed on delivery apps operating in your area, check out the restaurant partner portals for Doordash, Foodora, Uber Eats, and Postmates [8].

(4) Why is this innovative?

In short, it’s not. All sorts of foodservice operations—from mom and pop diners to international chains—batch their processing to save time and money.

Whether you’re turning 50 cows into 500kg of ground beef [9] or pre-peeling potatoes for easier mashing, dicing, and french-frying on the line; it’s easier to work in bulk:

Large franchises like McDonald’s and Pizza Hut [10] outsource their processing to local or national vendors;
Restaurant groups like Momofuku, Toronto’s Drake Hotel [11], and other hotels, chain restaurants, and catering companies set up an internal “menu” from which each unit orders common base ingredients (like sautéed onions); and
Individual kitchens schedule time for prep before service every day, so the cooks on the line have “ammunition” they can use during service.

These operations are often called commissary kitchens—though I cannot find a canonical source defining the term. It references the at-cost (sometimes cost-plus) stores you will find on military bases [12].

Like “ghost” vs “cloud” kitchens above, a commissary kitchen can also mean a shared kitchen with space for rent. These commissary kitchens are popular with food trucks and caterers [13], who use the space to—you guessed it—prepare base ingredients before service.

INSIGHT VS. INNOVATION: in my opinion, the ghost kitchen model is insightful, not innovative. It rewards operators who lean into the disintermediated customer relationship, offering restaurant group-like economies while you identify and iterate on promising concepts.

(5) Why should I be worried?

CONSUMERS: as we’ve seen, the ghost kitchen model rewards operators running a portfolio of menus. You’d hope this means a variety of cuisines, but just as often you find several “restaurants” with nearly-identical menus [14].

Ghost kitchens are a by-product of our changing consumption habits as our relationships with food become more transactional than cultural. In that vein, they’re no different from meal alternatives like Soylent or delivery kits like Blue Apron.

These changes put the onus on individual consumers to consume consciously; a company does not have your wellness in mind the same way a family member or trusted peer does.

As our world increasingly offers what Alex Danco calls “cooking as a service” [15], and opaque algorithms serve up the menus we order from, consumers should worry about who is choosing what they eat.

DELIVERY RESTAURANTS: first and foremost—if you’re already selling on a delivery app, and your kitchen has capacity, and you’ve not yet tested alternative menu concepts… you should start. Work with your cooks to create menus, with an aim to reuse food you might be “wasting” today. There are profitable concepts to be mined; and they will appreciate the autonomy and growth opportunities.

You should also be paying close attention to the delivery apps and cloud kitchen providers. For example, in a recent article for New Food Economy, journalist Corey Mintz heard this from a filmmaker working on a documentary about the delivery app ecosystem:

“If I had to guess where they’re going with this, at some point they collect enough data where they can effectively do the same thing Amazon did to its sellers, and cannibalize the best-selling items by making it themselves” —Matt Newberg [16]

NON-DELIVERY RESTAURANTS: while delivery demand is growing, it’s only one channel through which to build a successful restaurant.

If you are not yet offering delivery through the apps—if at all—I’d worry most about FOMO-driven decisions. Though it’s probably still worth keeping a watchful eye on delivery demand trends!

I hope you found this article on ghost kitchens informative and entertaining!

If you have any questions or feedback, or run a ghost kitchen yourself, please do reach out. I want to hear from you :)