Goal
To increase average product holding through personalised cross-sell marketing campaigns informed by a Next Best Action engine.
Approach
Carried out Sprint Zero and identified cross sell as the main opportunity area starting with Car to Home before expanding to other journeys.
Adapted WL’s playbook cross-sell template to specific portco’s infrastructure in order to develop production-grade code from the start.
Processed a dataset of 176m+ customers-month combinations and developed a predictive model that was able to identify 73% of cross-sell events in a month.
Leveraging tech assets, we expanded the initial cross-sell model from 1 to 9 journeys across 3 products using a recommender model that identifies what is the Next Best Action for a given customer.
Interventions
The workflow generates purchase likelihood estimates for 3-5 million customers every week. A series of filters select high propensity customers to be targeted.
The team worked with the marketing team to create personalised cross-sell campaigns. This new approach replaces the generic multi-product newsletter, previously used as the main vehicle to drive cross-sell.
The marketing campaigns delivered 70-80% increase in cross-sell performance across journeys. This represents ~$18m of incremental EBITDA.
The team up-skilled the internal data science team on technical best practices through day-to-day collaboration and on-the-job coaching.