How much does unexplained variance actually cost per year?

Every hospitality business has unexplained variance. It is the portion of the stock take gap that cannot be attributed to a specific cause: not a delivery shortfall, not a logged wastage event, not a known preparation loss. It is product that was expected to be there and is not, with no explanation attached.

Most operators treat unexplained variance as background noise. A cost of doing business. Something to note in the report and move on from.

This is a mistake, not because unexplained variance can always be eliminated, but because it is rarely as small as it feels. And because the cost of accepting it accumulates in a way that is invisible until you do the arithmetic.

What unexplained variance includes

Unexplained variance is a category of last resort. It is what is left after every other explanation has been applied.

Explained variance might include: a delivery that arrived short and was credited (supplier shortfall), over-preparation that produced more waste than the recipe assumed (prep variance), a portion that ran larger than specified during a busy service (portion drift), or staff consumption that was logged and accounted for.

What remains after these explanations is unexplained. It might be petty theft by a staff member. It might be a guest who helped themselves to something they should not have. It might be product that was used but not logged, because the process for logging was not followed consistently. It might be measurement error in the stock count itself.

The cause matters for the response. But before you know the cause, you need to know the scale.

The arithmetic

Consider a bar operation turning £8,000 per week in drinks revenue, with a drinks cost of around 28% (GP of 72%). The weekly cost of goods consumed should be approximately £2,240.

Unexplained variance at 3% of drinks cost — a figure that would not raise immediate concern on a weekly report — is £67 per week.

Over fifty-two weeks, that is £3,484.

At 5% of drinks cost, which is still within the range that many operations would describe as "acceptable," the annual cost is £5,824.

These are not dramatic numbers at the level of a single week. They are meaningful numbers at the level of a year, and they compound with scale: a group of five sites with the same variance rate is losing £17,000 to £29,000 per year to unexplained consumption.

Why the cost is usually higher than the variance percentage suggests

Unexplained variance tends to be underreported rather than overreported, for two reasons.

First, stock counts done under time pressure produce estimated figures rather than accurate ones. Estimates tend to be optimistic: the person counting has a mental model of what should be there and the count reflects that model rather than the physical reality. This systematic optimism suppresses the reported variance figure whilst the actual variance is higher.

Second, categories with high transaction volume and moderate unit value are harder to track accurately than categories with low volume and high value. A spirit that moves fifty units per week across dozens of transactions is harder to account for precisely than a premium whisky that moves two. The variance in high-volume, moderate-value products tends to be absorbed into the noise of the overall count.

The real cost of the manual counting process that produces these estimates compounds the problem further: estimated figures entered through a paper-to-spreadsheet process carry two layers of error before they are ever reviewed.

The response that the arithmetic demands

The purpose of naming the annual cost of unexplained variance is not to produce alarm. It is to make the investment in reducing it feel proportionate.

A business spending three to five hours per month improving the quality of stock counts, investigating specific variance categories, and implementing process changes that reduce unexplained loss can expect to reduce unexplained variance meaningfully. If the annual cost of that variance is £5,000 to £10,000, the investment in addressing it pays back quickly.

The operators who consistently run low unexplained variance are not doing something extraordinary. They are doing the ordinary things reliably: accurate counts, variance review before submission, prompt credit notes for delivery shortfalls, and a process for escalating when a specific category shows variance that cannot be explained by normal operations.

Variance categories and what each one tells you

Not all unexplained variance requires the same response.

A consistent negative variance in a specific product category, across multiple consecutive stock periods, suggests a systematic cause: a measurement issue in the count, a recurring delivery shortfall from a specific supplier, or consumption that is happening regularly without being recorded.

A sudden spike in variance following a specific event, a new staff hire, a change in the bar team, a period of unusually high volume, is worth investigating in the context of that event rather than as a standalone observation.

Variance that is randomly distributed across categories, small in each, and shows no pattern is most likely measurement noise from the counting process. Improving count accuracy is the appropriate response. Understanding whether variance is explainable or not is always the first question to answer before deciding how to respond.

What Hops does with variance data

Hops does not just report variance. The system analyses patterns across your historical data and, where consistent anomalies appear, surfaces them as prompts to investigate.

A product that consistently shows negative variance on weekend stock periods, for example, is flagged as a pattern worth examining: it might be a measurement issue in the Sunday count, or it might be something that needs a different conversation.

The principle is the same one that underlies the broader Hops approach to data: variance is not bad. Unexplained variance is bad. Understanding where your variance is coming from is the starting point for doing something about it.