Identifying the Base Sales Contribution for MTA Model

Identifying the Base Sales Contribution for MTA Model

Sandnya Dalvie, Sameer Siddiqui, Neelima Ratra

Abstract:

Multi-touch attribution (MTA) is a technique for measuring marketing effort that considers all the consumer touchpoints and assigns a certain percentage of credit to each channel so that marketers can evaluate the importance of each touchpoint in generating a potential sale.  MTA is a bottom-up strategy that uses cookies to give users'-level perceptions of every channel, format, and ad creative before the sale event.

The strategy underlines the fact that not all purchases are a direct result of marketing efforts but can also arise from other non-obvious factors when capturing customer event logs to attribute sales. This highly sophisticated automated solution considers factors like seasonality, specific macroeconomic factors (like change in economic index, inflation, or incentive to promote sales like lower loan interest rates), government subsidies to avoid falling in the pitfall of assigning all the credit entirely to Digital Marketing drivers. MTA then assigns sales credit to each of the touch points using a highly effective Deep Learning algorithm.

The goal of the article is to build a statistical process for determining which sales are driven by actual marketing efforts and which sales are generated without spending much on product advertisements, campaigns, or other non-promotional activities. It provides a detailed explanation of a variety of scenarios for predicting the Base Sales percentage for online leads and sale produced by modelling approaches.