Every claim about a promotion’s effect is a comparison with something that did not happen. The baseline is that counterfactual: the units the product would have sold in those stores, in those weeks, at its regular price and without the display or the feature. It is usually built from recent non-promoted weeks, adjusted for season, trend, distribution changes and other events that would have happened anyway.
Because the baseline is an estimate, the choice of estimate decides the result. Set it low — from a weak season, or by ignoring the dip after an event when shoppers who stocked up buy less — and every promotion looks profitable. Set it high and good promotions look like failures. A baseline recalculated after the event, by whoever is presenting the result, is the usual source of that bias.
Fixing the baseline method before the promotion runs, and applying the same method to every event, is what makes results comparable from one promotion to the next.