Rethinking Last Touch Attribution Data driven attribution modelling at scale
Marks & Spencer wanted to better understand the value of their display activity alongside their other digital activity. Incubeta was tasked with built an interactive attribution dashboard that would allow for easy comparisons between their existing attribution solutions and the off-the-shelf solutions provided by their web analytics platform.
We developed an attribution model that utilized a Markov simulation methodology allowing us to better estimate the revenue contributions of each marketing channel in a robust and scalable way. This transparent modeling approach provided additional insights to the M&S team beyond the black box nature of many data driven attribution solutions.
We utilized our in-house expertise to breakdown top line attribution results by channel, campaign & site, as well as both sales & marketing dates. From here we ingested 300+ GBs of log-level data from the client’s independent ad server, using Google BigQuery & Cloud Functions to analyse this data at scale.
This extensive ETL process stitched together impressions data alongside clicks, and session level data allowing us to build out the log of customer touchpoints required for our analysis.