Attribution Model

An attribution model defines how the success of a conversion is split across the marketing touches involved. Since users usually take several steps before buying, the model decides which channel earns credit for what. The choice of model changes the rating of every campaign and therefore also the budget.

Common models

Classic options are first click, last click, linear, time decay and position based. Data driven models derive the weights from actual path data rather than fixed rules. The latter need enough data and clean tracking, otherwise they get noisy.

Where to be careful

Models are not truth, they are agreements. Pure last click attribution almost always undervalues reach and brand channels. Data driven scoring needs consistent tracking across all channels, otherwise distortions appear. The important thing is to use one model consistently rather than switching all the time.

Relevance for JD Mail and JD Partner

In the JD context, sending data from JD Mail and click and reach data from JD Partner are important inputs for attribution. With tracking parameters passed through consistently, mail clicks, ad contacts and lead flows can be analysed together. You then see which mix actually sells instead of only crediting the last touch.