How to track liquor variance and find where the liquor goes

Your pour costThe cost of a drink's ingredients divided by its menu price, as a percentage. tells you how much of your sales went toward the liquor you used. Liquor variance helps explain why that cost might be higher than expected. If your counts show $175 of liquor went out each week that your sales and records can't explain, that's worth investigating: about $9,100 over a year. To find it, you need two numbers for each product: counted usage and sold usage.

What liquor variance measures

Counted usage comes from your counts: it is what left the shelf between two counts. Sold usage comes from your sales: each drink sold on the POSPoint-of-sale system: the register or software that records every sale., multiplied by how much of the product its specA drink's exact recipe: each ingredient and how much of it. pours. Liquor also leaves the shelf in other recorded ways: transferred to another bar or the kitchen, poured into house batches, or logged as waste and as compsA drink given away for free, for example to a regular or to make up for a mistake. that never went through the POS. Account for each of those separately. What is left is the variance. It might be a real loss, or it might be a mistake in the records.

Counted usage = last count + received + made in house − this count

Sold usage = sum of (drinks sold × what the drink's spec pours of this product)

Variance = counted usage − sold usage − poured into batches − transfers out − logged waste and comps

In the trade, counted usage is called actual usage and sold usage is called theoretical usage, so the method is called AvT (actual vs theoretical).

How to work it out, product by product

Take your house gin, in 1 liter (1 L) bottles. Your last invoice says $30 a bottle before tax. You count every Sunday after close.

You can check the 1-bottle gap from the shelf: 4 + 18 − 10.5 − 8 − 0.5 = 3. You should have found 3 bottles this Sunday, but you found 2. The 10.5 bottles of sold usage are worked out below.

Sold usage is most of the work. Your POS product mix report shows how many of each drink you sold. Find every drink that uses the gin, and multiply the number sold by its spec. This week you sold 150 Gin & Tonics with 2 oz (60 ml) of gin each, and 50 Negronis with 1 oz (30 ml) each:

150 × 60 ml + 50 × 30 ml = 10,500 ml = 10.5 bottles of 1 L

The batch line matters. If you leave out the 8 bottles that went into the freezer martini, the gin looks 9 bottles short instead of 1 (20 − 10.5 − 0.5 = 9). Count the batch left in the freezer like any bottle. If it is not on the count sheet, it reads as loss. When you count the batch as its own product, enter the gin poured into it on two lines. On the gin, enter it as "poured into batches". On the batch, enter it as "made in house", just as you enter a delivery.

Choose one way to account for the batch and use it consistently. If you track the freezer martini as its own product, record the gin poured into it and count the remaining batch separately. If you count the batch's remaining ingredients as gin-equivalent stock instead, don't also count them a second time as a separate batch.

Other mistakes also put a false gap into the numbers:

• Old specs. Use the specs your bartenders pour now. An old spec puts an error into every drink sold.
• Open POS buttons. Every POS button must point at a drink. An open "cocktail $14" button can sell gin that sold usage never sees.
• The wrong nights. Sales must cover exactly the nights between the two counts.
• Transfers. Record bottles sent to another bar or to the kitchen as transfers out, on their own line, or they read as loss.

Liquor variance percentage

Variance % = variance ÷ counted usage × 100

For the gin, that is 1 ÷ 20 × 100 = 5%. Then turn bottles into dollars: multiply the bottles unaccounted for by the price per bottle.

What the gap looks like in dollars

Run the same sums for every product you count. This week, counted usage across all products came to $7,000, and $175 of it was unaccounted for. That is 2.5% of counted usage.

Four products make up $155 of the $175, so start with them. Rank by dollars, not bottles, to decide what you look at first: the gin and the Chartreuse are 1 bottle each, but the gin is $30 and the Chartreuse is $60.

Value every bottle on the same basis for the period, such as the last invoice price, so a supplier price change doesn't distort the ranking.

The same gap shows in your pour cost. Sales this week were $35,000 before tax, and you used $7,000 of liquor, so your pour cost was 20%. Without the unaccounted $175, it would be 19.5%. So the $175 adds half a point to your pour cost.

What is a normal liquor variance?

Start by measuring your own bar. Track each product over several comparable counts, then look for changes that are larger than your usual counting differences.

For the whole bar, you can also turn unaccounted liquor into points of pour cost. In this example, $175 of unaccounted liquor against $35,000 of sales adds half a point to your pour cost. That's worth watching, but there is no industry-wide acceptable limit for it. Your goal is to explain the difference and reduce the losses you can prevent.

Robert Plotkin, writing in Bar & Restaurant, uses pour cost between counts as a warning sign: costs "typically shouldn't deviate more than a point between inventory periods". That is about how far your pour cost moves from one count to the next, not a limit on unaccounted liquor.

Where the gap comes from

Let the pattern of your variance tell you where to look first. Then rule out the possible causes one at a time.

Now take the Aperol from the dollar table. Your Aperol Spritz is 2 oz (60 ml) Aperol and 3 oz (90 ml) prosecco, and you sold 100 this week.

If the whole Spritz were poured too big, the Aperol and the prosecco would be short by a similar share. But the Aperol is 20% short and the prosecco only 4%. That suggests the problem may be specific to the Aperol rather than the whole Spritz being poured too generously. Start by checking the count, deliveries, other drinks that use Aperol and any unrecorded comps. Then run a pour test on the Spritz: each bartender pours the spec's measure at service pace into the glass you serve it in, and you measure each pour against the spec. If the numbers still don't reconcile, look at how the bottle is handled and stored.

Variance and bar shrinkage

Bar shrinkage is the real loss left after you take out errors in the counts, the deliveries and the sales. It covers over-pouring, spills, breakage, comps nobody recorded, and theft.

Shrinkage = variance − the part traced to count, delivery and sales errors

The variance tells you which product is short, not who took it, and a gap isn't proof that liquor was stolen. So rule out the other causes in the table above first. Conclude theft only when those checks find nothing.

Running it every count

Work out the variance at every count, for every product you counted both times.

Start with the products at the top of the dollar table. This week, those are the Chartreuse, the Aperol, the gin and the bourbon. Recount them the next morning before service, and check their deliveries and any batch they go into. The pour test is different: run it on the product short by the most bottles. This week, that is the Aperol.

One unusual count isn't enough to justify changing a spec or a long-standing process. Recount the product and check the records first. If the same unexplained gap keeps appearing, look at the pouring method, training and stock controls. But don't wait for three counts to correct an obvious problem, such as an unrecorded transfer, a faulty pour spout or missing stock.

A wrong count gives a wrong variance, so keep your count routine the same every week. This free inventory spreadsheet does the sums for you once you enter what you sold.

Questions

Sources

1. Cheryl S. Stanley, Alcoholic Beverage Costing Practices in the Hospitality Industry. Literature review, pp. 20 to 21: summarizes the beverage cost control systems in Keiser and DeMicco (1993), including inventory or quantity control and the standard-cost system.
2. Robert Plotkin, Control Pour Costs and Watch Profits Rise. April 10, 2012. How far pour cost should move between inventory periods; eight causes of a rising pour cost, including transfers, and the advice to rule them out by elimination.
3. Jeffrey Morgenthaler, How to Take Inventory and Calculate Pour Cost. April 10, 2014. Beginning inventory + purchases − ending inventory.
4. Institute of Licensed Trade Stock Auditors (ILTSA), How to prepare for a stock audit. A reconciliation against the register that lets losses be quantified and allocated to particular products.