A ghost kitchen sells through delivery apps from a kitchen with no dining room, often running several brands from the same line. The plan turns on one number: what an order keeps after the app's commission. A $24 order that keeps $13.94 when the customer picks it up keeps $8.90 through an app at the example's 25% commission, so the mix of app orders, your own website and catering decides whether the kitchen makes money. The other numbers are the usual ones: food cost, packaging, rent, wages and the cash it takes to get through the slow first months. The plan below works in channels per $100 of sales. In the example, a Miami ghost kitchen with three brands opens with $108,000, ends its first year about $12,000 behind before the owner's pay while orders build, and makes about $3,250 a month at full speed. Below is each part of the plan, the first year month by month, the startup budget, and a free template.
The short version
- Describe the kitchen in a paragraph: the delivery area, the brands, who orders, the goals.
- Work out what one order keeps through each channel: app, your own website, pickup, catering.
- Plan sales by channel per $100, with the app's commission in the variable cost.
- More than one brand from one line shares the rent, the staff and the inventory.
- Budget enough working cash for the months before orders build; check it against the lowest cash point.
- Move regulars to your own website and catering, where an order keeps more.
- TaroStack keeps one stock for all three brands, with each brand's recipes drawing on it. More below.
Part 1: the business description
Brickell Night Kitchen, the made-up ghost kitchen in Miami, Florida, I use in these examples, writes:
Brickell Night Kitchen is a delivery-only kitchen in a 900-square-foot second-generation restaurant space, running three brands on the delivery apps, wings, rice bowls and desserts, from one licensed kitchen, with its own ordering website for pickup and local delivery, and office catering. In its first year it aims to reach $47,000 of sales a month.
Part 2: what one order keeps
| One $24 order | Food | Packaging | Commission or fees | Own driver | Keeps | Share |
|---|---|---|---|---|---|---|
| Through a delivery app, 25% commission | $7.70 | $1.40 | $6.00 | $8.90 | 37.1% | |
| Own website, own driver | $7.70 | $1.40 | $0.96 | $3.00 | $10.94 | 45.6% |
| Own website, pickup | $7.70 | $1.40 | $0.96 | $13.94 | 58.1% |
The commission is the example's contract; yours is in your agreement with each app, and promotions inside the app cost extra. The website's 4% covers online ordering and card fees. A pickup order keeps more than half again what an app order keeps, which is why the kitchen puts a card in every app bag with a code for its own site. How a cost per order is built is in plate cost formula.
Part 3: customers and prices
The three brands share one line and one stock: chicken wings and thighs go into the wings and the bowls, rice into the bowls, and the dessert brand uses the fryer between rushes. Each brand's menu is short, so the line can cook any order in under ten minutes. The app menus are priced 10% above the website menu, to cover part of the commission. The license side, one food service license for the kitchen and a registered name for each brand, is in ghost kitchen license.
Part 4: what each channel earns
| Channel | Sales | Variable cost | Contribution | Margin | Hundreds a month at full speed |
|---|---|---|---|---|---|
| Wings brand, delivery apps | $100.00 | $63.00 | $37.00 | 37.0% | 180 |
| Rice bowl brand, delivery apps | $100.00 | $60.00 | $40.00 | 40.0% | 150 |
| Dessert brand, delivery apps | $100.00 | $55.00 | $45.00 | 45.0% | 50 |
| Own website, pickup and own driver | $100.00 | $48.00 | $52.00 | 52.0% | 60 |
| Catering orders | $100.00 | $41.00 | $59.00 | 59.0% | 30 |
Variable costs include food, packaging, the app's commission or the website's fees, and the kitchen's own drivers. At full speed the kitchen sells $47,000 a month and keeps $19,800, 42.1%; the apps bring 81% of the sales and $14,910 of what's kept. Every $100 that moves from an app to the website keeps $7 to $15 more.
Part 5: the first year, month by month
Fixed costs are $16,550 every month: rent ($3,500), wages for the line ($9,500), utilities ($1,200), insurance ($400), tablets and software ($250), promotions on the apps ($1,200), and other costs ($500). Sales start at 45% of full speed:
| Month | Sales | Profit | Cash, year to date |
|---|---|---|---|
| 1 | $21,300 | −$7,568 | −$7,568 |
| 2 | $24,500 | −$6,226 | −$13,794 |
| 3 | $28,200 | −$4,670 | −$18,464 |
| 4 | $31,600 | −$3,243 | −$21,707 |
| 5 | $34,700 | −$1,938 | −$23,645 |
| 6 | $37,600 | −$710 | −$24,355 |
| 7 | $40,100 | $352 | −$24,003 |
| 8 | $41,900 | $1,104 | −$22,899 |
| 9 | $43,800 | $1,908 | −$20,991 |
| 10 | $45,200 | $2,498 | −$18,493 |
| 11 | $47,000 | $3,250 | −$15,243 |
| 12 | $47,000 | $3,250 | −$11,993 |
The year's sales are $442,900, and the kitchen ends it $11,993 behind, before the owner's pay. It breaks even at about $39,300 of sales a month, in the seventh month. The first draft of the plan budgeted $20,000 of working cash; the template showed the lowest cash point at −$24,355 in month 6, $4,355 short, so the budget went up to $25,000.
Part 6: the startup budget
| Item | Example |
|---|---|
| Build-out of a second-generation kitchen | $30,000 |
| Fryers, rice cookers and reach-ins | $25,000 |
| Hood and fire suppression repairs | $8,000 |
| Lease deposit and first month | $7,000 |
| Opening food and packaging | $6,000 |
| Photos and brand design | $4,000 |
| Licenses and permits | $1,500 |
| Tablets and printer | $1,500 |
| Working cash | $25,000 |
| Total to start | $108,000 |
The $83,000 before working cash is earned back in a little over two years at full speed. Renting a station in someone else's kitchen instead is compared in ghost kitchen vs commissary kitchen.
Get the template
Download the ghost kitchen business plan workbook. Four sheets: your channels, each per $100 of sales with its variable cost and hundreds a month at full speed; the first year, month by month, with sales ramping up, profit, cash, the first profitable month, the lowest cash point and whether the working cash covers it; your fixed costs; and the startup budget. It's an ordinary Excel file with no macros, and it works in Google Sheets, LibreOffice and Numbers.
Where this stops working
Three brands on three apps each have their own menu, but they all take chicken, rice, oil and boxes from the same walk-in. When the wings brand has a big Friday, the rice bowls run out of chicken on Saturday, and no single app shows what the kitchen as a whole used.
How TaroStack does it
TaroStack keeps one stock for all three brands, with each brand's recipes drawing on it.
One stock, many recipes. Every menu item is a recipe by weight, and each prep batch, a tray of marinated wings or a pot of rice, takes from the same stock, so the kitchen's food cost is one number.
Counts with no signal. The walk-in can be counted on a phone with no signal, so Saturday's chicken is known before the rush.
Reorders. The reorder list shows chicken, rice and boxes in days of cover, across all three brands.
| The job | Three app dashboards | In TaroStack |
|---|---|---|
| Chicken used this week | Added up by hand | Prep batches from one stock |
| Cost of a rice bowl | A guess | True cost from the recipe |
| Boxes before Friday | Counted on Thursday | The reorder list |
Recipes, counts and the reorder list are on every plan, from $49 a month. The first 30 days are free, with no credit card. If you'd like a hand setting up, ask, and we'll do it with you.