The four points of food cost nobody is looking at

When food cost drifts, the first instinct is almost always the menu. Reprice the steak, drop the lobster, put a cheaper fish on. It is visible, it feels decisive, and in our experience it recovers about half a point — while quietly costing more in average spend than it saves.

The money is somewhere less interesting. Across twenty-two audits in the last three years, four leaks appear again and again, in properties from forty covers to four hundred rooms. None of them require a new menu. Together they are worth roughly four points.

1. Purchasing has drifted

Specifications were written once, at opening, and have not been checked since. Suppliers have substituted grades, pack sizes have changed, and the price comparison everyone believes is happening stopped happening when the purchasing manager left. In a third of audits we find the same product being bought from two suppliers at a 19% price difference, in the same week.

2. Payroll is modelled on covers you no longer do

Rotas ossify. A property that has lost its lunch trade will often still be staffing for it three years later, because the rota is inherited rather than built. This is a payroll problem that presents as a food cost problem, because the same brigade produces more waste when it has less to do.

3. Nobody owns the yield

Recipe cards say 180g. Nobody has weighed what actually leaves the pass in eighteen months. The gap between specified and served portion is the single largest recoverable item we find, and it is almost never deliberate — it is what happens when a section is under pressure and no one has been shown the number.

4. The menu mix has not been read

Not the menu — the mix. Which items sell, which carry contribution, and which do neither but stay on because someone likes them. Most operations have this data and nobody opens it.

3.8

Average points of food cost recovered

22

Audits in the dataset

12 wks

Typical time to hold the gain

0

New menus required

“Every operator I meet believes their problem is unique. It is almost never unique. It is almost always one of these four.”

None of this is difficult work. It is unglamorous, it requires someone to sit with the numbers for two weeks, and it produces a result that shows up in the P&L rather than in the dining room. That is precisely why it does not get done.