We address statistical issues in attributing revenue to marketing channels and inferring theudimportance of individual channels in customer journeys towards an online purchase. We describeudthe relevant data structures and introduce an example. We suggest an asymmetric bathtub shapeudas appropriate for time-weighted revenue attribution to the customer journey, provide anudalgorithm, and illustrate the method. We suggest a modification to this method when there isudindependent information available on the relative values of the channels. To infer channeludimportance, we employ sequential data analysis ideas and restrict to data which ends in audpurchase. We propose metrics for source, intermediary, and destination channels based on twoandudthree-step transitions in fragments of the customer journey. We comment on theudpracticalities of formal hypothesis testing. We illustrate the ideas and computations using dataudfrom a major UK online retailer. Finally, we compare the revenue attributions suggested by theudmethods in this paper with several common attribution methods.
展开▼
