What is a ghost kitchen, and how does it make money?

A ghost kitchen is a kitchen with no storefront. Orders arrive through delivery apps or a website, and the food leaves in a bag. DoorDash's merchant guide describes it as a facility built only to fill delivery orders, with no dining room and no walk-in guests.

Cloud kitchen and dark kitchen mean the same place. Nation's Restaurant News said in 2019 that ghost and cloud kitchens are close to interchangeable, and that the language was still settling.

A virtual restaurant is different. It is the brand and the menu, not the building. It can be cooked in a ghost kitchen, a shared kitchen or an existing restaurant's kitchen. Think of the kitchen as the place and the virtual restaurant as the concept.

A guest finds a brand on an app, orders, and a ticket appears on a tablet. The cook makes it and bags it. A courier collects it at the door.

There is no front of house. That removes servers, a dining room and the rent on a busy street. It adds new work: packaging that survives a ride, photos that sell, and a menu short enough to cook fast.

One kitchen can serve several brands. DoorDash says operators may run three, five or more than ten concepts from one kitchen, each with its own menu and name.

Most ghost kitchens fit one of these shapes. Each trades control for cost.

• Rent a station in a shared ghost kitchen building. You share space with other brands, and DoorDash's guide notes that peak-hour scheduling can be tight.
• Rent from an operator-managed kitchen. The operator does more, and DoorDash's guide says these usually mean less control and higher monthly fees.
• Add a delivery-only brand to a restaurant you already run. NRN calls this a virtual brand. It uses a kitchen you already pay for.
• Lease your own commercial kitchen and run it for delivery only. You control the layout and carry the lease.

A ghost kitchen earns money like any kitchen: the price, minus food, packaging, labour, rent and fees. The difference is that fees are larger and rent is smaller.

Here is one order, as an example and not a forecast. A guest pays $30 on a delivery app. Food costs 30%, or $9, inside the 28% to 35% range a Johnson & Wales University pricing article gives for profitable restaurants. Packaging is $1. An app plan at 25% takes $7.50. That leaves $12.50 before labour, rent and equipment.

At $12.50 an order, $2,000 of monthly rent alone needs 160 orders to cover. Run that math before you sign anything.

On your own site, the same dish pays card processing instead of a marketplace cut. Stripe charges 2.9% plus CA$0.30 on domestic cards in Canada, about $1.20 on this order. You also pay your own driver, so count that cost.

Estimates vary a lot. In 2020, NRN reported a CBRE figure of as little as $20,000 to open one, against millions for a brick-and-mortar restaurant in New York City. DoorDash's 2026 guide gives a higher range. It lists $75,000 to $200,000 to start, $1,000 to $3,000 a month for shared kitchen space, and $15,000 to $25,000 for kitchen-only equipment.

These are the guides' own estimates, and DoorDash is a delivery app. Treat them as ranges to test, not quotes.

Costs the guides do not price for you: packaging, insurance, your own labour, app promotions, and any deposit or term on the kitchen. Some operators also take a share of orders. NRN reported in 2020 that two large operators charged a flat monthly fee plus 15% of orders. Ask for the full price list in writing.

A ghost kitchen is still a food business. The rules come from your local health authority, not from the app. FDA tells new food businesses to contact the local health department, and says state and local governments regulate restaurants and food trucks. The US Small Business Administration says local governments set registration, licensing and permit rules.

In Toronto, anyone who intends to operate a food premises must notify the Medical Officer of Health before starting. The city reviews plans for new premises, and posts inspection results online on DineSafe. It says to check zoning and building rules before you sign a lease.

Shared kitchens add a question: who answers for what? In a 2020 NRN report, New York City's health department said the lease owner is responsible for violations in shared spaces. Ask the kitchen's owner for its health permit, and ask your health department whether you need your own.

The apps bring orders. They also bring four problems.

• The fee. In Canada, Uber Eats lists 20%, 25% or 30% on delivery orders, depending on the plan. DoorDash lists 20%, 25% or 29%.
• The customer. Check what each app tells you about the guest. If you cannot email them yourself, you do not own the relationship.
• The menu. DoorDash's own guide warns that editing menus and prices by hand across platforms causes errors and lost orders.
• The listing. You do not control how an app shows your brand next to the next one.

Both apps also sell direct ordering. DoorDash says orders through its Online Ordering are commission-free, with payment processing fees only. Uber Eats lists 2.9% for its webshop. Compare the terms you would sign, including what you give up.

The practical move is a second channel you own. Put a link on a card in every bag. Count how many orders arrive direct each week.

Any online ordering tool needs the same basics. That means a menu with prices, hours and limits, card payments, a ticket that reaches the kitchen, and a way to deliver. Our guide to setting up online ordering walks through each one.

pickup.kitchen is one way. Each brand can have its own site and menu, managed from one account. It adds a kitchen screen, stock counts and delivery tools for your own drivers. It charges 10% per order with no monthly fee, and on card orders the guest pays the 10% at checkout. It does not import orders from the apps, so you will run two systems.

Work through these in order. Each one costs less to check now than to fix later.

• Pick a menu that travels well and cooks fast. Cost every dish: food, packaging and your time.
• Call your local health department before you sign a lease or a kitchen contract.
• Ask the kitchen operator for its permit, its fee list in writing, and its rules for peak hours.
• Write your break-even: monthly rent and fixed costs, divided by what one order leaves after food, packaging and fees.
• Check zoning and any business licence with your city.
• Plan who delivers. Apps bring couriers. Your own site needs your own drivers.
• Photograph every dish in daylight and name it the way guests search for it.
• Open one brand first. Add more only when the first one runs smoothly.
• Track orders each week by channel, app and direct.

It suits a cook with a delivery-friendly menu and tight costs. It suits a restaurant that wants a second brand without a second lease. It suits less well food that loses quality in a bag, or a business that needs a dining room to sell.

If you are unsure, test cheaply. Run a delivery-only brand from a kitchen you already have for a month, and read the numbers.

Whatever you choose, keep one number on the wall: how many orders each channel brought this week. It tells you where the next hour of effort pays.