How much does a promotion really sell?

The report says the promotion sold €14,000. But part of it would have sold anyway. This is how new sales are separated from everything else.

• Sales during a promotion are not the effect of the promotion.
• The effect is measured against a scenario without the promotion, estimated from your history.
• What would have sold anyway, purchases brought forward and cannibalisation all have to be taken out.

What the report does not tell you

When a promotion ends, the report usually shows the sales for the period: €14,000 during the discount week, for example. It is a real figure, but it answers a different question. It does not say how much you sold because of the promotion, only how much you sold while it was running.

The difference is not a detail. A promotion mixes customers who were going to buy anyway, customers who brought a purchase forward, sales taken from other products in your store and, finally, genuinely new sales. Only the last one is the effect of the promotion.

The right question: what would have happened without it?

To measure the effect, you compare what happened with what would have happened without the promotion. That second scenario cannot be observed, because the promotion did run, so it has to be estimated. In causal analysis it is called the counterfactual.

Comparing with the week before looks simple but easily goes wrong: the season, payday or a price change on another product all move sales on their own. A good counterfactual takes all of that into account, from each product's history.

An example with numbers

Back to the €14,000. Suppose that without the promotion the product would have sold €6,100 on those same dates. Another €1,900 are purchases brought forward: customers who would have bought the following week. And €1,400 came from similar products in your own catalogue. What is left, €4,600, is the new sales the discount created.

With that figure, and with what the discount cost across everything sold, you can tell whether the promotion made money. The conclusion can change completely: an offer that looked like a success by volume may leave no margin, and a quieter one may be among the most profitable. The figures in this example are illustrative.

What to do with it

Measuring every past promotion this way gives you a map of what works in your business: which products respond to a discount, at what depth and on which dates. That map is the basis for planning the next ones.

Effecta does this calculation with your own data, promotion by promotion: before each launch, to decide, and afterwards, to measure what happened.