Investigate inventory variance

Actual usage = beginning inventory + received + transfers in − transfers out − ending inventory. Expected usage is the sum of sold menu-item quantities multiplied by their recipe quantities, adjusted only for documented events that your policy treats separately.

Usage variance = actual usage − expected usage − documented waste. A positive result means measured usage exceeded recipe-based usage after the listed adjustment; it does not identify why. inventory variance, actual usage, and projected usage are the three terms in that sentence.

Resolve scope and timing problems before interpreting behavior. A small setup error repeated across many sales can create a large apparent variance.

1. Confirm the opening and closing counts cover the same products and locations.
2. Check deliveries, transfers, returns, and count cutoff times.
3. Verify unit conversions and container sizes.
4. Review recipe quantities and POS item mappings.
5. Confirm documented waste, comps, and batch preparation.
6. Only then investigate repeated operational patterns.

Sort exceptions by dollar impact. A 10% miss on well vodka usually matters more than a 40% miss on a slow amaro. Coach from money, not from embarrassment.

Check recipe maps and POS item names before you accuse a shift. Unmapped cocktails deplete stock that projected usage never sees. That looks like theft and is often a setup error.

Repeated SKUs, repeated shifts, or repeated stations are the only patterns worth a people conversation. One noisy week after a menu launch is not a verdict.

ExactPour reason codes (over-pour, breakage, spill, measurement, unknown) build a history. Use them. A pile of unknown codes means the team is not finishing the investigation, not that the math failed.

Visual fill estimates, missing transactions, substitutions, evaporation, and recipe drift can all affect the result. Use variance as an investigation queue, not as evidence about a person or a cause.