Key takeaways
- A ghost kitchen is a place that cooks only for delivery or pickup; a virtual brand is a menu that exists only online, often cooked in an existing restaurant.
- The appeal is incremental orders from spare kitchen capacity, but marketplace fees still apply to every order.
- Virtual brands fail for predictable reasons: thin margins after fees, kitchen strain at peak and weak guest loyalty.
- All the normal rules still apply: health permits, allergen information and, for larger chains, menu labeling.
The terms, untangled
"Ghost kitchen," "dark kitchen," "virtual brand" and "delivery-only restaurant" are used loosely and often interchangeably. They describe different things, and the difference matters when you decide whether any of them fit your business.
| Term | What it is | Who usually runs it |
|---|---|---|
| Ghost kitchen (or dark kitchen) | A kitchen that cooks only for delivery and pickup, with no dining room | Operators renting space in a shared facility, or a single brand in a low-rent location |
| Virtual brand | A restaurant brand that exists only online, with its own name and menu, cooked in a kitchen that also serves another brand | Existing restaurants using spare capacity, or companies licensing brands to restaurants |
| Delivery-only restaurant | A real, single brand that sells only off-premises | Independent operators and chains testing a market |
| Shared or commissary kitchen | A licensed kitchen rented by the hour or month to several businesses | Caterers, food trucks, small brands and ghost kitchen tenants |
| Licensed virtual brand | A ready-made brand and menu that a restaurant cooks under license from the brand owner | Restaurants that want a virtual brand without developing one |
For most independent restaurants, the realistic option is a virtual brand cooked in their own kitchen. Renting a ghost kitchen is a different decision, closer to opening a new restaurant without a dining room.
Why the idea took hold
Off-premises dining is now the majority of restaurant traffic: nearly 75% of US restaurant traffic happens off-premises[1]. Younger guests drive much of it. 51% of Gen Z and millennial adults say picking up takeout or ordering drive-thru is essential to their lifestyle, and 41% say they rely heavily on delivery[2].
Marketplaces made it easy to reach those guests. Third-party delivery is available at 58% of limited-service and 32% of full-service restaurants[3], and a new listing can appear in front of thousands of app users without a storefront. Restaurants with quiet kitchens between rushes saw a chance to sell more from the same space, staff and inventory. The growth in limited-service restaurants, which added 3,249 establishments in 2025[4], reflects how much of the market is built around speed and off-premises orders.
How a virtual brand runs from an existing kitchen
A virtual brand works by sharing almost everything with the host restaurant except the name, menu and listing.
Build the menu from existing inventory
The best virtual brand menus use ingredients the kitchen already stocks, so there is little extra purchasing or waste. A burger restaurant might run a chicken sandwich brand from the same fryer and buns.
Create a separate listing
The brand gets its own name, logo, photos and descriptions on one or more delivery apps, and sometimes its own ordering page.
Route orders clearly
Tickets for the virtual brand must be clearly marked so the line knows which brand and packaging to use. If several apps and brands run at once, integration matters.
Package and label it separately
Its own stickers or bags, so the guest receives what they ordered from the brand they chose.
Fit it into the kitchen's rhythm
Many operators limit virtual brand hours to periods with spare capacity, such as late afternoon or late night.
The kitchen's address, permits and staff stay the same. To the guest, the virtual brand looks like a separate restaurant on the app.
The economics
The case for a virtual brand is incremental margin: extra orders that use spare capacity, where rent, equipment and some labor are already paid for. The case against is that every one of those orders carries marketplace fees, and those fees are high relative to restaurant margins. Two in three restaurants say their average fee to delivery platforms is between 15% and 29.9% of the transaction[5], and in a May 2026 survey 47% of full-service and 43% of limited-service operators said they make no profit on third-party delivery orders[6].
Example: a fictional pizza restaurant launches a wings brand from its own kitchen. It sells 25 orders a day at an average of $24. After a 25% commission, packaging and the food and direct labor for the wings, it keeps about $4 an order before overhead, or about $100 a day. That looks worthwhile, until the restaurant notices that wing orders peak on Friday nights, when the fryer is needed for its own menu, and pizza ticket times rise. The incremental profit is real only if the virtual brand's orders land when the kitchen has room.
Licensed brands add another layer. If you cook a brand under license, the brand owner usually takes a fee or share of sales on top of the marketplace fee, and may set the menu, ingredients and suppliers. Read those terms as carefully as a franchise agreement.
Why many virtual brands do not last
Virtual brands fail for predictable reasons, and most are visible before launch.
- Margins disappear after fees. A brand that needs discounts to sell is paying for orders twice.
- The kitchen is strained at peak. Virtual orders arrive when the host restaurant is busiest, and quality suffers on both brands.
- No loyalty. Guests who find a virtual brand on an app rarely know or remember it, and there is no dining room to build a relationship.
- Too many similar brands. Popular categories fill with near-identical virtual listings, and ranking on the app becomes a paid contest.
- Weak listings. Generic stock photos and thin descriptions convert poorly.
- Mixed quality. A menu built quickly from spare ingredients may not match what guests expect from a brand that looks specialized.
- Platform and guest pushback. Guests who discover that several app listings come from one kitchen may feel misled, and platforms periodically change policies on duplicate menus and virtual listings.
None of these are reasons never to try. They are reasons to test cheaply, measure honestly and stop quickly if it is not working. Our checklist on whether to launch a virtual brand turns them into decisions.
The rules still apply
A virtual brand does not escape any of the rules that apply to the kitchen it is cooked in.
- Health permits and inspections cover the kitchen, whatever brands it sells under. Check whether your local health department needs to be told about additional brands.
- Allergen information must be accurate for every brand's menu, and cross-contact risks are higher when several menus share stations.
- Menu labeling: FDA calorie labeling requirements apply to restaurants and similar retail food establishments that are part of a chain with 20 or more locations[7]. If your group is that size, check how the rules apply to your virtual brands' menus.
- Business registration: a trading name may need to be registered locally before you sell under it.
- Platform policies: read each marketplace's rules on virtual brands, duplicate menus and the information you must display about the kitchen.
Being open about the kitchen behind the brand, for example "From the kitchen at [your restaurant name]" in the store description, is good practice and protects you if guests or platforms ask.
Renting a ghost kitchen instead
Some operators skip the host restaurant and rent space in a ghost kitchen facility: a building divided into small licensed kitchens, often with shared receiving, dishwashing and a pickup counter for drivers. It can be a way to enter a new neighborhood without the cost of a dining room, or to run a delivery-heavy concept separately from a full-service restaurant.
The trade-offs are different from a virtual brand. Rent, equipment and staff are new costs rather than shared ones, so the location has to generate enough orders on its own to cover them. Everything depends on marketplace reach and your own online marketing, because there is no foot traffic. Facility contracts vary widely: ask what is included, such as utilities, cleaning, storage and driver handoff, what the minimum term is and how the facility handles orders for many tenants at once.
Treat a ghost kitchen like any new location. Build a forecast of orders per day, test the menu's travel quality, model margin after all fees and rent, and decide in advance what results would make you stay or leave at the end of the first term.
Signs a virtual brand is working
After two or three months, look at five measures: orders per day, margin per order after all fees, the share of orders that come at times the kitchen has spare capacity, refunds and ratings, and ticket times for the host restaurant during virtual brand hours. A brand that is profitable, keeps its rating and does not slow the host restaurant is worth keeping. One that fails on two or more measures usually does not improve with time.
Operationally, the same disciplines that make any delivery work apply here: clear order routing, accurate packing and a well-run handoff. See taming tablet chaos for running several apps and brands from one kitchen, and delivery app listings that sell for building a listing guests trust.
Frequently asked questions
What is the difference between a ghost kitchen and a virtual brand?
A ghost kitchen is a physical kitchen that cooks only for delivery and pickup, with no dining room. A virtual brand is an online-only restaurant brand, which can be cooked in a ghost kitchen or, more often, in an existing restaurant's kitchen.
Are virtual restaurant brands profitable?
Some are, when they use spare kitchen capacity, share ingredients with the host menu and are priced to cover marketplace fees. Many are not, because fees are high, orders clash with the host restaurant's peak and guests show little loyalty.
Do I need a separate permit for a virtual brand?
The kitchen's existing permits usually cover the food prepared there, but rules vary. Check with your local health department and register any new trading name as your locality requires.
Can I run a virtual brand without delivery apps?
You can sell through your own ordering page, but most virtual brands rely on marketplace discovery because the brand has no storefront or existing guests. Without an app, you need your own way to reach new customers.
References
National Restaurant Association (2025). From Trend to Transformation: Off-Premises Dining Now Essential for Restaurant Consumers, Operators (opens in a new tab) ↑
National Restaurant Association (2025). Report: Takeout, drive-thru, delivery are more popular than ever (opens in a new tab) ↑
National Restaurant Association (2026). National Restaurant Association comments to the FTC on the Food Delivery Fees ANPRM (Project No. P267101) (opens in a new tab) ↑
US Bureau of Labor Statistics (2025). Quarterly Census of Employment and Wages, 2025 annual averages, NAICS 722513 (US private) (opens in a new tab) ↑
National Restaurant Association (2026). National Restaurant Association comments to the FTC on the Food Delivery Fees ANPRM (Project No. P267101) (opens in a new tab) ↑
National Restaurant Association (2026). National Restaurant Association comments to the FTC on the Food Delivery Fees ANPRM (Project No. P267101) (opens in a new tab) ↑
US Food and Drug Administration (2026). Menu Labeling Requirements (opens in a new tab) ↑