Cost Per Serving: The Only Grocery Number Worth Tracking
What cost per serving actually is
Cost per serving is the register price of a recipe's ingredients divided by the number of servings the recipe makes. Not the price of the amounts the recipe calls for - the price of the packages you have to buy to cover those amounts, because that is what the checkout charges you.
That distinction is the whole reason this number is worth tracking. Recipes talk in amounts: 6 ounces of pasta shells, 3/4 cup of pumpkin, 2 pork chops. Stores talk in packages: a 12-ounce box, a 15-ounce can, a 2-pound tray. The gap between those two languages is where most published "this dinner costs $6" claims fall apart. When we priced 712 dinner recipes item by item at a real Kroger store, the honest math started with real packages every time - and this post explains that method so you can run it yourself on any recipe. It is the foundation for everything else in our Meal Planning series.
A worked example: 12 servings from a $24.92 basket
Take pumpkin ricotta stuffed shells, a 12-serving USDA recipe we priced at a Kroger in July 2026. Its eight non-pantry ingredients came to $24.92 at the register. Divide by 12 servings: $2.08 per serving, which lands it in the cheapest quarter of everything we have priced.
The basket is ordinary stuff: a $2.99 box of jumbo shells, a $2.99 tub of ricotta, a $2.99 can of pumpkin, a $1.99 jar of pasta sauce, $4.99 of grated parmesan, plus garlic, fresh basil, and a shaker of sage. No single item over five dollars. The recipe's power is not cheap ingredients - it is that a normal basket gets divided by twelve plates instead of two or four. Yield is doing the heavy lifting, and yield is exactly what a receipt never shows you.
Why the receipt total lies to you
A cheaper receipt can be the worse deal. In our data, orange pork chops - two chops, a sweet potato, half an orange - rang up at just $15.24. But the recipe makes 2 servings, so that is $7.62 per serving: almost four times the stuffed shells, on a receipt that is $9.68 smaller.
| Recipe | Basket total | Servings | Cost per serving |
|---|---|---|---|
| Pumpkin ricotta stuffed shells | $24.92 | 12 | $2.08 |
| Beef pot roast | $42.63 | 8 | $5.33 |
| Orange pork chops | $15.24 | 2 | $7.62 |
Source: CentsPerServing corpus pricing, Kroger store 01400513, July 2026. Register-rounded package prices, pantry staples excluded.
The pork chops are not a bad dinner, and $15.24 is not a lot of money. But if you are choosing recipes by what the trip costs, you will systematically pick small-yield dinners and wonder why the monthly grocery total keeps climbing. Per-serving cost is the number that survives the comparison. Why the small basket is so expensive per plate has a second cause too - package rounding - and that is next.
Package rounding: you cannot buy 1.3 cans
The register rounds every ingredient up to whole packages, and that changes real cost more than most people expect. A beef pot roast calling for 2.5 pounds of chuck meant buying three 1-pound packages at $10.99 each: $32.97, not the $27.48 that naive per-pound math predicts. One line item, $5.49 of rounding.
It happens on small items just as often. A lentil soup we priced called for one 28-ounce can of diced tomatoes; the cheapest diced tomatoes at that store come in 14.5-ounce cans at $1, so the basket holds two cans. Our matcher applies a small grace factor for recipe rounding - two 14.5-ounce cans (29 ounces) cover a "30 ounces of tomatoes" line, because no shopper buys a third can for one missing ounce. But the direction is always up: packages quantize your groceries, and the smaller the recipe, the more of the basket is rounding rather than food.
Why pantry staples are left out
Salt, pepper, cooking oil, and water are excluded from every per-serving cost we publish. The reason is fairness in both directions: billing a $3 salt canister to one dinner would make that recipe look absurdly expensive, while pretending oil is free would flatter deep-fried recipes. Staples amortize across dozens of meals, so the convention is to leave them out.
The line has to be drawn narrowly, though. A quarter teaspoon of allspice is not a pantry staple if you do not own allspice - and our pot roast basket honestly bills the full $3.99 jar for it. Same with one tablespoon of orange juice pulling a 46-ounce bottle into the cart at $4.49. The convention we settled on: only true universals (salt, pepper, oil, water) are excluded, and everything else gets counted at package price with the remainder treated as carryover. Optional garnish lines are also left out of totals, since skipping them changes nothing about the dish.
Remainders: the money that carries forward
Package rounding has a flip side: what you did not use is still in your kitchen. A beef stew we priced needed 6 ounces of bacon; the cheapest package was 16 ounces at $4.79. The basket paid for all 16, but 10 ounces of bacon are now sitting in the fridge waiting for a carbonara.
This is why cost per serving is a planning tool, not just a scorecard. If next week's menu is chosen to use this week's remainders, the second basket collapses. In our real-world test, week 1 of a planned menu cost $34.60 at the register - and week 2 cost only $13.91, because the plan pre-checked what week 1's packages had left behind before picking recipes. The weekly plan builder automates exactly that: shared packages get counted once across the week instead of once per recipe.
The first-basket effect: spice jars and big bottles
Some recipes look expensive per serving only because it is your first time buying their supporting cast. That pot roast at $5.33 per serving includes a whole jar of allspice and a whole bottle of orange juice - items that will still be there for the next ten dinners that want them.
A useful way to read a priced basket, then: separate the meat-and-produce core from the durable tail of jars, bottles, and sauces. The core is the recipe's true recurring cost; the tail is a one-time buy-in that amortizes. First baskets overstate a recipe's long-run cost, sometimes by a couple of dollars per serving. If a recipe you want to repeat looks pricey, check whether the expensive lines are things you will ever have to buy again. Repeat cooking is cheaper than the first basket suggests - one more reason a single receipt is a bad verdict on a recipe.
What counts as a good cost per serving
Benchmarks make the number useful. Across the 712 dinner recipes we fully priced at a real Kroger store, the median was $4.62 per serving, and half of all recipes landed between roughly $3.16 and $6.89. Under $3.16 is the cheapest quarter; under $2 is the cheapest 7% - 48 recipes, almost all built on beans, lentils, and rice.
So a practical scale for a home-cooked dinner: under $3 per serving is genuinely cheap, $3 to $5 is normal, $5 to $8 is on the expensive side, and past $8 you are into seafood and specialty territory. The full distribution, including the $0.41 floor and the $46.84 ceiling, is in our breakdown of what dinner actually costs in 2026, and the What It Costs series keeps publishing priced cuts of the same data. Wherever your own recipes land, the point is the comparison: a number only means something next to other numbers.
How to run the math on your own recipe
You can compute cost per serving manually in four steps. First, cross off pantry staples: salt, pepper, oil, water. Second, for every remaining ingredient, write down the package you would actually buy - not the amount the recipe wants - rounding up to whole cans, bags, and pounds. Third, total those package prices. Fourth, divide by the recipe's declared servings.
Two honest warnings from doing this 712 times. Watch for count-versus-weight traps: "4 chicken breasts" is roughly 2 pounds, and pricing it as one package undercounts. And trust the yield printed on the recipe only if it looks sane - a "serves 2" recipe with a pound of meat in it is really a serves-3 with leftovers, and the per-serving number moves accordingly. If you would rather not do any of this by hand, the recipe pricer runs the same method automatically: paste a recipe URL or ingredient list and it matches each line to real products at your local Kroger, does the package rounding, and flags any match it is not sure about. On a hand-graded evaluation, about 93% of its ingredient matches were shopper-accepted.
One number, most of the leverage
Cost per serving will not clip coupons for you, but it quietly drives the decisions that matter: which recipes earn a weeknight slot, whether a batch is worth doubling, and whether a "budget dinner" headline is telling the truth. Track it and the grocery bill follows.
Start with the recipes you already cook on repeat. Price them once, honestly, with package rounding included. Most people find their rotation spans a wider range than they guessed - often $2 to $8 per serving inside the same household - and shifting two dinners a week toward the cheap half moves a monthly bill by real money without changing how you eat. Everything else on the blog builds on this one habit.
Get the number without the spreadsheet
Paste a recipe URL or ingredient list and get its priced basket and cost per serving at your store - free, no account.
Price a recipeFAQ
What does cost per serving mean?
Cost per serving is what the groceries for a recipe cost at the register, divided by the number of servings the recipe makes. It is the only grocery number that lets you compare a 12-serving casserole against a 2-serving skillet dinner on equal terms.
How do you calculate cost per serving?
Price every non-pantry ingredient as the packages you would actually buy - whole cans, whole pounds, rounded up - add them into a basket total, then divide by the recipe's declared servings. A $24.92 basket for a 12-serving recipe is $2.08 per serving.
Is there a free cost per serving calculator?
Yes. The recipe pricer at /price takes any recipe URL or pasted ingredient list, matches each ingredient to real products at your local Kroger, does the package rounding, and returns the basket total and cost per serving - free, no account, three baskets a day.
What is a good cost per serving for dinner?
Across 712 dinner recipes we priced item by item at a real Kroger store in July 2026, the median was $4.62 per serving. Under $3.16 puts a recipe in the cheapest quarter, and under $2 puts it in the cheapest 7% - mostly bean and rice territory.
Why is my real cost higher than the recipe site claims?
Most published recipe costs use per-ounce math on exactly the amount called for. Stores sell packages, not amounts: 2.5 pounds of chuck roast means three 1-pound packages, and a 28-ounce can of tomatoes may only exist as two 14.5-ounce cans. Register prices round up.
Do pantry staples count in cost per serving?
By convention, no. Salt, pepper, cooking oil, and water are excluded because they amortize across dozens of meals - billing a whole jar to one dinner would distort the comparison. Everything else, including spices you may need to buy, gets counted.
What about the leftover ingredients a recipe leaves behind?
Package remainders - the 10 ounces of bacon left from a 16-ounce package - are real value that carries into future meals. Basket totals include them, which is why planning the next week around remainders is powerful: one of our real weeks cost $34.60 and the next only $13.91.
Why compare recipes per serving instead of by basket total?
Because yield varies wildly. In our data, a $24.92 basket made 12 servings ($2.08 each) while a $15.24 basket made 2 servings ($7.62 each). The cheaper receipt was almost four times more expensive per plate. Basket totals hide exactly the number that matters.
Data: CentsPerServing corpus pricing runs, July 2026 - item-level shelf prices with package-count math at Kroger store 01400513, from a 1,651-recipe corpus with 712 fully matched dinner recipes per store. Worked examples (pumpkin ricotta stuffed shells $24.92/12 servings, orange pork chops $15.24/2, beef pot roast $42.63/8) are individual priced baskets from that dataset; store prices change without notice. Ingredient matching is ~93% shopper-accepted on a hand-graded evaluation, and uncertain matches are flagged. This article reports our own measurements and is not financial advice.