Café food cost: does the 30% rule work for resold pastries?

Not item by item. The 30% (or 33%) food cost line is a target for everything you sell together, not a markup every item has to hit. A croissant you buy ready-made already carries the bakery's labor in its price, so its cost % runs higher while your own payroll stays lower. Give resold items their own target, price them from their delivered unit cost, and check that your blended food and drink cost and your prime cost still sit inside your range.

Where the 30% and 33% come from

Both numbers describe a whole business: all the food and drink you used in a month, divided by all your sales. The National Restaurant Association's figures for 2024 put the median cost of food and non-alcoholic drinks at 32.4% of sales for limited-service restaurants (counter service) and 32.0% for full service. For limited service, it notes that its 2010, 2013 and 2016 reports averaged about 33%. The Association also says plainly that these figures aren't meant as standards or goals for any one restaurant. They're medians: about half of the restaurants that answered ran higher, and half lower.

RestaurantDoctorAI's targets work the same way. A counter-service café falls under quick service, where our range for food and drink cost combined is 26–32% of sales, and prime cost (food and drink cost plus total labor, divided by sales) has a ceiling of 60%. Both are for the whole café, across everything on the menu. The food cost guide has the ranges for every type of restaurant.

So if you mark every item up three times, you're treating a whole-menu average as a rule for each item. Three times cost is a 33% cost on every item. That works for some items and prices others out of reach.

What a made item costs compared with a bought one

When you make something in-house, its food cost counts only the ingredients. The person who made it is paid out of your labor line. When you buy something ready-made, the price you pay already covers the maker's labor, ovens and margin. Nothing about it shows up in your payroll, but all of it shows up in your food cost.

The IRS draws the same line for taxes. Its Tax Guide for Small Business says a merchant counts the cost of the goods it bought for sale, at the prices actually paid rather than the list price, and that freight on those goods is part of cost of goods sold. It adds that labor usually belongs in cost of goods sold only for manufacturers, and that small retailers usually have none to put there. How you file is a question for your accountant. For pricing, the takeaway is that a resold item's cost is what you paid to have it on your counter, delivery included.

That flips if your baker is on the clock anyway with spare hours: then the made croissant's labor isn't extra, and it leaves $3.47. The point isn't that buying beats baking. It's that you can't compare the two on food cost %, because one has its labor inside the price and the other doesn't.

A resold item isn't labor-free either. Someone receives the delivery, fills the case, warms, bags and throws out what didn't sell. It's usually seconds per item rather than the hours it takes to bake, but it's there.

Why one blended number can mislead you

Because each kind of item runs at its own cost %, your blended number moves when the mix moves, even if every price is right. Here's the example café's month, split by what it makes and what it resells.

The pastries run at 42% and the bottled drinks at 50%, well above any 30% rule. Yet the whole café sits at 30.1%, inside the 26–32% quick-service range, because the coffee drinks run low. With $8,700 of labor (29% of sales), prime cost is $17,740 ÷ $30,000, or 59.1%, just under the 60% ceiling.

Now say a new office nearby adds $3,000 a month of pastry sales, at the same 42% and with the same staff hours.

Judged by the blended number alone, that month looks like a problem. It isn't one. The reverse happens too. Say sandwich cost creeps from 32% to 38% (an extra $360 a month) in a month when coffee sales grow by $2,000 at 22%. The whole café moves from 30.1% to 30.8%, less than a point, and nothing on the P&L tells you it's the sandwiches. A target per category catches it the month it starts.

Set a target for each category

Split your menu into the same groups your POS already reports, then sort them into made and resold. Give each group a target range of its own. The ranges below are the example café's own choices, not a standard. They follow its own costs.

Then test the targets against your mix. If every category ran at the top of its range, the example café would spend $4,000 + $2,040 + $2,250 + $1,500 = $9,790 on $30,000 of sales: 32.6%, just over its range. So it keeps the coffee drinks near the middle of theirs, where most of its sales are. Your targets only work if, weighted by what you actually sell, they land inside your whole-café range and keep prime cost under your ceiling.

To see each category's real cost, you need two splits each month:

• Sales by category from your POS category report, before tax and tips.
• Purchases by category from your invoices. Ask your bookkeeper to give resold goods their own accounts (bakery items, bottled drinks), apart from coffee, milk and sandwich food. Count each group in your month-end inventory too, so the cost is what you used, not just what you bought.

How to price a resold item from its unit cost

Start with what one item costs you to sell, not the case price. Three things go in: the delivered price of one, the share you throw out unsold, and the packaging that goes out with each one.

Dividing by the share you sell matters more for pastries than for bottled drinks. A sealed bottle can wait on the shelf until it sells. A croissant has a day, and each one you throw out raises the cost of the ones you sell.

When a high markup is costing you sales

A higher markup leaves more per item, but only on the items that sell. If guests pass on a pastry at its price, the dollars per item don't help. To compare two prices, work out how many you'd need to sell at the lower one to take in the same dollars after cost:

Say the almond croissant sells 120 a month at $7.25. That leaves 120 × $4.85 = $582. At $5.25, each leaves $2.85, so the café needs $582 ÷ $2.85 = 204.2, or 205 a month, to take in as much. If it sells more than 205 at the lower price, the lower price leaves more. If it sells fewer, the higher one does.

Nothing here tells you which way your guests will go. A test does. Change one item's price, keep everything else the same, and count it in your POS item report for four weeks against the four weeks before. For raising prices on the items you make, see how much to raise menu prices, and for costing a cup, see what cafés can do about coffee prices.

Check the whole café each month

Category targets are how you price. The whole-café numbers are how you check. Once a month:

• Each category against its own target. A category over its range for two months running is the one to look at: re-cost its best sellers on this month's invoices.
• Blended food and drink cost against your range. If it moved but every category held, your mix moved. Check the dollars left before you act on it.
• Prime cost against your ceiling. It's the one number that counts both the labor a made item takes and the labor a resold item doesn't.

Many counter-service restaurants run above the 60% line. The National Restaurant Association's 2025 Restaurant Operations Data Abstract put the median limited-service restaurant's prime cost at 65 cents of every sales dollar. For how to work out yours, see the prime cost guide.

This week

1. 01Split last month's sales. From your POS category report, sort every category into made in-house or resold, and write down each one's sales.
2. 02Split last month's purchases the same way. Sort the invoices, or ask your bookkeeper for a separate account for resold goods. Work out each category's cost %.
3. 03Cost your three best-selling resold items. Delivered unit cost ÷ the share you sell, plus the bag, in the plate cost calculator. Write down the cost per one sold.
4. 04Pick one item priced at three times cost that isn't selling. Work out how many you'd need to sell at a lower price to match, then decide whether to test it for four weeks.