How to Calculate Restaurant Wine Cost (and Why Recorded Cost Isn’t COGS)

The core formulas

1. Wine cost of goods used (COGS) = beginning inventory value + purchases − ending inventory value.
2. Wine cost percentage = wine COGS ÷ net wine sales × 100.
3. Variance = actual usage cost − recorded (theoretical) cost of what the POS says you sold.

‘Net wine sales’ should mean the same thing every period: wine revenue after discounts and comps, excluding tax and service charges. Decide your definition once and keep it.

Recorded cost versus actual COGS

Recorded cost is what your system believes the wine you sold should have cost: each bottle or glass sold multiplied by its unit cost. Actual COGS comes from the counts and invoices — what physically left inventory, whatever the reason.

The two are never identical. Recorded cost leaves out unrecorded comps, tastings, spillage, breakage, over-pouring and transfers. Actual COGS includes all of them. When actual cost is higher than recorded cost, the difference is the money to explain. A cost percentage alone tells you very little; the variance tells you where to look.

A step-by-step monthly process

1. Count all locations on the last day of the period, by SKU and vintage, including open BTG bottles.
2. Value the count at current invoice cost per bottle, after distributor discounts and credits.
3. Total purchases from invoices received in the period — not orders placed, not payments made.
4. Record adjustments: kitchen transfers, credits for breakage or corked bottles, and any transfers between outlets.
5. Calculate usage COGS and the cost percentage against net wine sales for the same dates.
6. Run recorded cost from the POS sales mix and compare it to actual usage cost.
7. Investigate the largest variances by dollar value and write down the cause.

Worked example (hypothetical numbers)

Adjustments that distort the number

• Kitchen wine: transfer it out at cost so it lands in food cost, not wine cost.
• Comps and manager tastings: record them so they appear as an explained category, not a hidden variance.
• Credits: post distributor credits in the period they apply, or the month looks worse than it was.
• Vintage and price changes: a new vintage at a higher cost changes your inventory value even if bottle counts are steady.
• Large purchases near month-end raise ending inventory and can make one month look unusually good and the next unusually bad. Look at several periods together.

What wine cost percentage should we target?

There is no universal number worth copying. The right target depends on your concept, list depth, price points, BTG share, storage commitments and how much aged inventory you deliberately hold. A program carrying a deep cellar will look different from one buying weekly. Set your target from your own consistent data, then manage the variance between recorded and actual cost — that is what you can control month to month.

Common mistakes

• Using purchases as cost. Buying is not using; this ignores inventory change entirely.
• Mixing tax or service charges into sales.
• Valuing counts at list price instead of actual invoice cost.
• Comparing a month with a large cellar purchase to one without, and drawing conclusions from a single period.