Sales rise sharply during a promotion and fall below normal afterwards. The second effect is as important as the first and is frequently left out of the evaluation.

Three sources make up the increase

Extra volume during a promotion comes from shoppers switching from a competing product, from shoppers buying earlier or in greater quantity than planned, and from genuinely new consumption.

Only the last is unambiguously additional. Switching may or may not persist, and buying forward is a timing change that reverses in the following period.

Separating the three requires comparing against what would have happened without the promotion, which cannot be observed directly and has to be estimated.

The dip is the buy-forward reversing

Shoppers who stocked up do not return to the category on their usual schedule, so the weeks after a promotion show sales below the normal baseline.

The depth and length of that dip depend on how storable the product is and how much was bought forward, and for durable goods it can run for months.

Measuring only the promotional period therefore overstates the result, sometimes by enough to turn a losing promotion into an apparent success.

Baselines are estimated, not measured

Assessing a promotion requires an estimate of the sales that would have occurred anyway, built from historic patterns adjusted for season and any other activity running at the time.

The estimate is sensitive to its assumptions, and small changes in how the baseline is drawn can materially change whether a promotion appears profitable.

Holding a comparable set of locations or customers out of the promotion gives a cleaner reading, which is why controlled tests are the preferred method where they are practical.

Category effects can outweigh product effects

A promotion that moves volume from one product to another within the same retailer adds nothing overall, even though the promoted line performs strongly.

Conversely, a promotion that brings shoppers into the shop can add value through the rest of their basket while losing money on the promoted item itself.

Evaluation therefore has to be set at the level the retailer actually cares about, and the answer frequently differs from the product-level result.

Why the practice persists regardless

Even where the arithmetic is unfavourable, promotions defend share, and stopping unilaterally cedes volume to competitors who continue.

They also serve internal purposes such as hitting period targets, clearing stock positions and meeting agreements with suppliers who fund part of the cost.

Understanding this explains why promotional intensity rarely falls even in categories where the measured incremental effect is known to be small.