Align dates and valuation first
Choose one start and end cutoff for inventory, purchases, transfers, and sales. Price theoretical recipes consistently with the period being investigated. A current purchase price compared against an old recipe price produces a variance even if staff follow the recipe exactly. Record that price effect before attributing the balance to operations.
Report currency and percentage points
In a fictional month, actual ingredient consumption is $19,600 and theoretical cost is $18,200 on $70,000 net food sales. The gap is $1,400. Actual food cost is 28%, theoretical is 26%, and the difference is 2 percentage points. The relative overrun against theoretical is $1,400 ÷ $18,200 = 7.69%; that is a different measure.
Build an explanation without counting a loss twice
Suppose invoice corrections explain $300 and documented unmodeled staff meals explain $400. The remaining unexplained gap is $700. A waste log may explain some of that remainder, but first check whether the same waste is already included in the recipe yield. Classify each adjustment once and preserve the original records.
Use the residual to choose an investigation
Recount expensive ingredients, check receiving quantities and pack conversions, then compare observed portions with recipe standards. Look for missing transfers or supplier credits before drawing conclusions about behavior. Repeat the same comparison next period after one process change. A narrowing unexplained gap is useful evidence; a single variance is not proof of waste, misconduct, or a specific cause.
Use your own operating data
Supplier prices, yields, labor rates, and channel fees change. KitchenMargin treats them as editable inputs rather than universal facts.