How a restaurant financial model fits together
A restaurant financial model is a set of linked tables. Startup costs say how much cash you need to open, assumptions say how many guests you expect and what you spend each month, menu costing says what each plate costs, and the projection and break-even tables turn those inputs into monthly figures. The links are the point: when one input changes, every figure that depends on it should change with it.
This article follows the numbers in order and shows how one workbook, the Restaurant 5-Year Plan & Menu Costing Workbook (https://northgateops.gumroad.com/l/restaurant-5-year-plan-menu-costing-workbook), arranges each part. It gives no benchmark ratios.
Startup costs and funding
Startup costs are everything you pay before the first sale: equipment, build-out, deposits, licences, opening stock and the like. A useful list is itemised and carries a contingency line for costs you did not think of. Whether the contingency counts as spent changes the cash left on opening day and how long payback takes.
The other half is funding: owner cash, a loan, other sources. Total costs minus total funding is the funding gap. A model that shows the gap on its own line makes an underfunded plan obvious before a projection is built on it.
In the workbook, the Startup costs sheet holds a pre-opening budget in seven categories and one contingency share of the subtotal, which it counts as spent and adds to the total. Funding goes on three funding lines labelled Owner cash, Loan and Other funding, and a Funding gap line shows the part of the total the funding does not cover.
Assumptions
Assumptions are the drivers behind every later number: opening date, seats, how many covers (guests served) you expect per day, the average check, days open, labour, rent, utilities, marketing, loan terms and growth rates. They belong on one sheet, because a rent figure typed in three places will be out of date in one of them.
Mark inputs visibly, so a reader can tell a figure you chose from a figure a formula produced, and treat any pre-filled default as a placeholder, not evidence of how your restaurant will look.
The workbook's Assumptions sheet uses named input cells for these drivers, plus a tax rate placeholder that no calculation reads. Yellow cells with blue text are inputs and grey cells with black text are formulas. The blank template starts with every input empty, apart from the 12 monthly seasonality factors, which start at 100% (no seasonality).
Menu costing and the sales-mix food cost %
Menu costing starts from an ingredient price list. Each ingredient has a unit, a purchase price, a quantity per purchase and a usable share, since trimming and cooking loss mean you do not serve everything you buy. Each recipe then lists ingredients and quantities, which gives a cost per portion. Divide that by the menu price and you have the item's food cost %. What a sale leaves after ingredient cost is its contribution margin.
One food cost % per item is not yet one for the restaurant. Items sell in different volumes, so the figure the projection needs is a weighted one: total recipe cost over total revenue, at the mix of portions you expect to sell. A plain average of item percentages would give a rarely ordered dish the same weight as the top seller.
In the workbook, the Ingredients sheet takes up to 300 ingredients and the Menu costing sheet takes up to 200 menu items with up to 12 recipe lines each. It shows cost per portion, food cost % and contribution margin per item, and a sales-mix weighted food cost %. Enter each item's portions sold in the Sales mix (portions sold) column, as a count or as a share of all portions sold (not revenue), using the same format for every item. Menu prices and the sales mix set the weighted food cost %, while revenue comes from covers times the average check typed on Assumptions, and the Break-even sheet shows the menu-implied average check beside the typed one.
The 60-month projection
A projection converts assumptions into months. Revenue is covers times the average check, usually with a ramp-up, because a new restaurant does not open at full volume. Cost of goods follows from the food cost % applied to that revenue. Labour, occupancy and other operating costs come next, and what remains is EBITDA, earnings before interest, tax, depreciation and amortisation. After that come loan interest and principal and a simple cash flow that starts from the opening cash left after startup costs. Sixty months is long enough to show a ramp-up and several seasons.
In the workbook, the 5-year projection sheet has 60 monthly columns and 5 annual columns, running from covers and average check with a ramp-up, through cost of goods, operating costs and EBITDA, to loan interest, principal and cash flow. Food cost % is one ratio across all 60 months, and the cost inflation input does not apply to it.
Break-even
Break-even asks how much you must sell to cover fixed costs. In its simplest form, fixed costs divided by the contribution margin on each unit of revenue gives break-even revenue, and dividing by the average check gives guests per day. Because contribution margin depends on food cost %, break-even is where menu costing and the projection meet.
A second question is timing: in which month do cumulative earnings reach the total spent on startup? The two answers differ, and a good model shows both.
In the workbook, the Break-even sheet shows break-even revenue, covers per open day and months to cumulative break-even, followed by a 5 by 5 sensitivity table over average check and food cost %. Months to cumulative break-even is the first month in which EBITDA added up since opening reaches the total startup costs. These figures are before loan payments and tax, so covering a loan payment from cash takes more than they show.
Checks and limits
Linked models fail quietly when an input is blank or a recipe points at an ingredient that is not on the price list. The workbook's Dashboard sheet shows one page of key outputs and an input check list, and a recipe line whose ingredient is missing from the Ingredients sheet is costed at zero and flagged, not silently ignored.
The workbook has no balance sheet and does not compute tax. It is a planning tool, not financial, tax, legal or lending advice, and it makes no promise about profit, returns or loan approval. No figure in either file is a benchmark or a forecast.
Excel, Google Sheets and the worked example
The download holds two .xlsx files, a blank template and a worked example filled in with invented numbers for an invented restaurant. The files have no macros and no links to other files. Formulas are restricted to a documented function allowlist (SUM, SUMIFS, SUMPRODUCT, INDEX, MATCH, IF, IFERROR, MIN, MAX, ROUND, EDATE, PMT, IPMT, PPMT and TEXT), and both workbooks use no other function.
You open an .xlsx file in Excel, and Google Sheets can import one, though an imported file may look different and formatting can shift. The files have not been tested in Excel and have not been tested in Google Sheets, so the documented function allowlist is the compatibility check the build can run, and it is not a test of either program.
Where to get the workbook
The Restaurant 5-Year Plan & Menu Costing Workbook is on Gumroad: https://northgateops.gumroad.com/l/restaurant-5-year-plan-menu-costing-workbook
This article was written with AI assistance, and the workbook it describes was built with AI assistance and generated by a script that has automated tests.