To manage food cost across locations, you need one source of truth for prices and recipes, a costed standard recipe for every menu item, and a weekly variance review that compares theoretical cost to actual cost per site. The locations that drift are the ones nobody is measuring, so the discipline lives in the weekly number, not the year-end number.
When I ran 14 locations putting out 8,000 meals a day, food cost was never a single problem. It was 14 problems wearing the same uniform. One site ran 31 percent, the next ran 38 percent, and both managers swore they were doing the same thing. They were not. You cannot fix what you cannot see, and most multi-site operators are flying blind on the only line item big enough to sink them.
Here is the math that should keep you up at night. On a group doing 40 million dollars a year in sales, a single point of food cost is 400,000 dollars. A 7-point spread between your best and worst site is not a rounding error. It is the down payment on a fifteenth location, and it is sitting in your walk-ins as waste, in your portions as over-pours, and in your invoices as creep nobody caught. The good news is that the same spread that costs you is the proof that a better number is achievable inside your own four walls.
Why Food Cost Drifts the Moment You Add a Second Location
A single restaurant is held together by the owner walking the line. You catch the over-portioned plate, the case of produce rotting in the back, the bartender free-pouring. Add a second location and you lose the one tool that worked: your own eyes. Now you are relying on people who portion differently, order differently, and define waste differently.
Drift is not a moral failure. It is a systems failure. Without a shared standard, every location invents its own, and the gap between your best site and your worst site becomes pure lost margin. I have seen two managers, both good people, both convinced they ran a tight kitchen, post numbers five points apart for a full quarter because nobody put the two reports next to each other on the same page.
- Different managers buy from different vendors at different prices for the same item
- Recipes live in someone's head, so portions vary plate to plate
- Waste gets thrown out, not logged, so it never shows up until inventory
- Nobody compares sites weekly, so a 5-point gap hides for a quarter
- New hires learn from whoever trained them, so bad habits replicate site to site
Build One Source of Truth Before You Touch Recipes
The first move is not a new recipe. It is a single, shared list of what every ingredient costs and what every dish should cost. When all 14 of my sites priced chicken differently in their own spreadsheets, comparison was impossible. Once every location pulled from the same costed item master, I could finally rank them and coach the laggards.
Theoretical food cost is what your recipes say you should spend at the volume you sold. Actual food cost is what your invoices and inventory say you actually spent. The gap between them is the entire game. Everything else - the recipes, the audits, the manager calls - exists to shrink that one number.
| Metric | Why it matters | Target |
|---|---|---|
| Theoretical food cost % | What recipes predict at actual sales | 28-32% |
| Actual food cost % | What invoices and counts prove | Within 2 pts of theoretical |
| Variance (actual minus theoretical) | Pure waste, theft, or over-portioning | Under 2 points |
| Inter-site spread | Best location vs worst location | Under 4 points |
Standardize the Recipe, Standardize the Plate
Once you can see the gaps, you close them with a costed recipe for every item: exact quantities, exact yields, exact plating. This is the playbook. A standard recipe is not about taking the chef's soul away. It is about making sure the burger in location 3 costs the same to make as the burger in location 11, so when one site's cost spikes, you know it is execution, not the menu.
I made portion scales and printed recipe cards non-negotiable. Within two periods, my worst site dropped from 38 percent to 33 percent with no menu changes at all. The food was already designed to hit margin. The site just was not following the design. The recipe card did not make anyone a better cook. It made the kitchen repeatable, which is a different and more valuable thing across 14 buildings.
A Worked Example: Finding the Five Points
Let me walk a real pattern. Site 7 posts a 37 percent actual food cost against a 31 percent theoretical. Six points of variance on a site doing 30,000 dollars a week is 1,800 dollars a week walking out the door, or roughly 94,000 dollars a year from one location. Here is how I would tear that apart in a single afternoon.
- Pull theoretical versus actual and confirm the six-point gap is real, not a counting error in last week's inventory
- Check receiving: are invoices being matched to the order, or is the site paying for short cases it never received
- Spot-weigh five center-of-plate portions during service against the recipe card to test for over-portioning
- Walk the walk-in and the trash for logged versus unlogged waste, especially produce and protein nearing date
- Reconcile the POS mix against depletion to flag comps, voids, and the quiet possibility of theft
Nine times out of ten the six points are not one big leak. They are two points of over-portioning, two points of unlogged waste, and two points of receiving sloppiness, each invisible on its own and obvious once you stack them. That is the entire reason the weekly number beats the year-end number: small leaks compound, and only a tight cadence catches them while they are still small.
The Most Common Mistake: Managing the Average
The trap I see most often is operators celebrating a group food cost of 32 percent and going back to sleep. The group average is a liar. It hides a 28 percent star carrying a 38 percent disaster. Averages let your worst site coast on the back of your best one, and your best manager eventually notices they are subsidizing someone else's slack and starts looking for the door. Never manage the average. Manage the spread, and manage the bottom two sites by name every single week.
Run the Weekly Variance Review
Annual food cost is a eulogy. Weekly food cost is a steering wheel. Every week, pull theoretical versus actual for each location, rank them worst to best, and put the bottom two on a 15-minute call. You are not hunting for blame. You are hunting for the one process that broke.
- Pull theoretical and actual cost per site every week
- Rank locations and flag any site over 2 points of variance
- Identify the single root cause: receiving, portioning, waste, or theft
- Assign one fix with an owner and a date
- Re-check the same site the following week
See exactly what waste, over-portioning, and shrink are costing you across every site before you build the fix.
Calculate your food waste costYour Weekly Food Cost Checklist
If you want a starting point you can run this Friday without buying anything, work this list. It is the same skeleton I used across all 14 sites, and it costs nothing but discipline.
- Every location pulls ingredient prices from one shared item-cost list, updated weekly
- Every menu item has a costed recipe card with exact yield and plating
- High-cost categories like protein and produce are counted weekly, not monthly
- Theoretical and actual food cost are reported per site on one page
- The bottom two sites by variance get a 15-minute root-cause call
- Every fix has one named owner and a recheck date the following week
The Bottom Line
Managing food cost across locations is not about being cheaper. It is about being consistent and visible. One source of truth tells you where you stand, standard recipes tell every site what good looks like, and the weekly review keeps drift from compounding. Start by getting all your locations onto the same costed item list this week. You cannot scale on spreadsheets and heroics, and food cost is where that truth shows up first.