Attributed sales versus incremental sales

Attributed sales are purchases credited to an advert under a reporting rule. Incremental sales are the additional purchases caused by the advertising compared with what would have happened without it. A shopper whose purchase is attributed might have bought anyway, so the two figures answer different questions.

What attribution counts

An attribution report needs a qualifying advert interaction and purchase. Ask which clicks or views qualify, how long a purchase remains eligible, which products and channels count, and how credit is assigned when several adverts qualify. Keep those rules with the reported amount.

The sales basis matters too. Gross sales can include orders later returned or cancelled; net sales deduct returned, cancelled or refunded orders. A figure for the promoted item and its variants covers a different product set from a wider same-brand, same-category figure. Label each amount accordingly.

Attribution can help monitor credited outcomes under stable rules or identify products and placements for investigation. It cannot reveal the purchases that would have happened without advertising.

What an incremental estimate needs

Define the outcome, such as eligible units or net sales, and compare the treated population with a credible estimate of its untreated outcome. A randomised holdout can provide a strong comparison when assignment, delivery and sales measurement are sound. Matched-market and modelled approaches depend on their own assumptions and data quality.

Ask how groups were assigned or matched, what sales and exposure data were covered, what else changed during the study, and how uncertainty was assessed. Stock shortages, promotions and different baseline demand can distort a weak comparison. A positive estimate with wide uncertainty is still uncertain.

A pre-campaign versus post-campaign sales change does not answer the same causal question on its own: demand may have changed for other reasons.

Explain a gap between the figures

An attribution report may credit purchases from shoppers who would have bought without the campaign. Its eligible customers, products, period and sales basis may also differ from those in an incrementality study. Compare the definitions before setting the results side by side; dividing one by the other is meaningful only when the populations and measures align and the resulting ratio answers a defined question.

Use attributed sales to describe credited purchases under the stated method. Use a credible incremental estimate, with its uncertainty and scope, to discuss additional sales caused by advertising. If no suitable causal study is available, report incrementality as unknown.