Sales being down 12% tells you what happened, but not why. The decline could reflect weaker demand, lost distribution, poor availability, a pricing change, less promotional support, a competitor launch or customers shifting into another part of the range.
The total alone cannot separate those explanations.
I usually start by breaking the result into a few simpler questions.
Did distribution change?
A product can lose sales while performing perfectly well in the stores where it remains available. If it has been removed from a retailer, lost shelf space, or gone out of stock across part of the network, the problem may be availability rather than demand.
Did rate of sale change?
If distribution is stable but sales are falling, I want to know whether each store is selling fewer units.
That points toward a different set of possible causes: price, packaging, competition, seasonality, promotional timing, or a change in customer behaviour.
Did the product mix change?
Revenue can grow while unit sales fall if customers move toward more expensive products. The reverse can happen when promotions or lower-priced products take a larger share.
A top-line percentage can hide a very different story underneath.
Is the category moving too?
A product declining inside a growing category is a different problem from a product declining because the entire market is shrinking.
The first may suggest lost share or a product-specific issue. The second may require a broader response.
The purpose of the first analysis is not to find the answer immediately. It is to narrow the field.
Sales down 12% is an alert.
The diagnosis begins when you work out whether the change came from distribution, demand, price, availability, mix, or the market around it.
Sales decline was only the alert. In the BIC case study, combining three datasets identified what customers could no longer understand about the product; the Lego range analysis applies the same discipline to a different commercial question, where the evidence describes changes in the range rather than proving a revenue cause. Both sit within the broader way I approach commercial analytics: narrow the problem, test competing explanations and stop the data claiming more than it can support.