Your sales report is lying to you, and it gets more convincing every week

Every planning system in retail learns from the same place: what sold. It is the most trusted number in the business. It reconciles to the till, it reconciles to finance, and nobody argues with it.

It is also, for a meaningful share of your range, wrong in a way that gets worse on its own.

How it happens

The following Sunday the store runs out again, a little earlier, and records 11. The forecast drifts down to 11. The week after, 10.

The system is now learning from a number it caused, and every cycle it becomes more confident about it.

Nothing in the process is broken. The forecast is accurate against recorded sales. The order matches the forecast. The store sells everything it receives. Every individual step checks out, and the business is steadily shrinking its own shelf.

Why the reports don't catch it

Availability is usually measured on whether there was stock at the time of the count, or on the days that had any stock at all. A line that sells out by 11 every morning shows as available. A category can look healthy on every dashboard while regular customers quietly learn not to rely on you for it.

Worse, the lines this hurts most are your fastest movers, because they are the ones most likely to sell through before the next delivery. The better a product sells, the more its history understates it.

What fixing it involves

The principle is simple: recorded sales during a stockout are not demand, so you reconstruct what demand would have been before you train anything on it. In practice that means knowing when the shelf went empty, not just whether it did, and estimating the missing sales from how comparable stores sold on the same day, how the line normally depletes through the day, and what customers substituted to.

None of this is new mathematics. It is well understood and has been for years. The reason it is so rarely done is that the report looks fine, so nobody goes looking.

How to check your own business

Take your top 200 lines by sales. For each store, count the days in the last quarter where closing stock was zero or close to it. Then look at the forecast for those lines and ask whether it has been drifting down while the category around it held steady.

If a line runs out regularly and its forecast is falling, you are almost certainly looking at this. In most businesses we look at, it is a handful of lines in every category, and they are rarely the ones anyone expects.