Restaurant Food Cost Control Across Multiple Locations

Every restaurant group tracks food cost. Most track one number, and one number cannot tell you anything actionable.

Knowing what percentage you spent on food tells you what happened. It does not tell you why, and it does not tell you what to do differently on Monday.

Restaurant food cost control needs a second number

The number missing from most groups is theoretical cost. What the food should have cost, based on everything that actually sold and what each dish is made of.

Put the two side by side and you get a variance. That variance is the whole point. It is the part of your spend that cannot be explained by what you sold, which means it was waste, over portioning, breakage, theft, unrecorded staff meals, or stock that walked out of the back door.

Without the theoretical number you are left looking at a percentage moving up and down and inventing stories about why.

Recipes are the price of entry

The theoretical number requires a costed recipe for every dish. There is no shortcut and this is where most efforts stall, because it is a lot of unglamorous work.

Two things make it manageable. First, start with volume. A minority of dishes usually make up most of your covers, and costing those gets you most of the value. Second, accept that recipes need maintaining. A recipe costed once and never revisited drifts as suppliers, pack sizes and specifications change, and a stale recipe produces a variance that is really a data problem.

Account for prep yield as well. A case of produce does not all reach the plate. If your recipe uses raw weight and your kitchen uses trimmed weight, your theoretical cost is wrong by the difference on every single dish.

Counts decide how useful the variance is

Variance is only meaningful over a period bounded by two stock counts. So the frequency of counting sets the resolution of everything else.

Count monthly and a variance tells you something went wrong somewhere in four weeks. That is too wide to investigate. Count weekly on the items that carry the value and the movement, and the variance points at a specific week, a specific site and often a specific category.

Consistency matters more than completeness. A count taken at the same time, in the same order, by people who count the same way, is worth more than an exhaustive count done differently each time. Inconsistent counting produces variance that is really measurement noise, and chasing noise burns credibility fast.

Compare sites fairly or nobody will listen

The obvious move with multiple locations is to rank them by food cost percentage. It is also the fastest way to lose the room.

Sites differ in menu mix, and dishes differ in margin. A site selling more of a low cost high margin dish will look better than a site doing the harder trade, regardless of how well either is run. Rank them raw and the managers who know this will discount the whole exercise.

The fair comparison is variance against each site own theoretical cost. That asks whether a site is running to its own achievable number, which is a question about execution rather than about which dishes happen to sell there.

Menu mix moves the number more than people expect

When group food cost moves and nobody changed anything, mix is usually the answer.

A promotion, a seasonal special, a weather change or a price adjustment shifts what sells, and what sells carries the cost. That is not a failure. It is a commercial outcome and it belongs in a different conversation from waste.

Separating mix effect from execution variance is what stops an operations team being held responsible for a marketing decision. It also tells the commercial team what their promotion really cost.

Purchase price is its own variance

The third moving part is what you paid. Supplier price changes, substitutions and off contract buying all shift cost without anybody in the kitchen doing anything differently.

Tracking purchase price variance separately keeps the conversation clean. Otherwise a supplier increase turns up as a kitchen problem, and the kitchen spends a week looking for waste that is not there.

This is also where a lot of quiet money sits. Off contract purchasing, where a site orders outside the agreed supplier or product, is common and rarely visible unless the system is set up to show it.

Make it weekly or it will not change anything

A food cost report that lands with month end is a history lesson. The manager who could have acted has already lost the month.

Weekly is the rhythm where this becomes a management tool. It requires faster counts, a POS feed that arrives daily and purchase data that is current. That is a systems investment, and it is the one that converts a reporting exercise into a margin change. Our guide to NetSuite for restaurants covers the underlying build.

Where to start

Cost the dishes that make up most of your volume. Set a weekly count on high value items. Produce theoretical against actual for one site for a month. Then take it to the manager and see whether the variance matches what they already suspected.

If it does, you have a working measure and you can extend it. If it does not, you have a data problem to fix before rolling anything out, which is a much cheaper thing to discover at one site than at thirty.

TechCloudPro works with restaurants and hospitality groups on the NetSuite work behind this and on the integrations that make weekly reporting possible.