• reubensegelbaum

Machine learning provides proven results


A large multi-national financial service provider had seen new enrollments for one of their flagship credit cards gradually decline for over a year, with channel data showing that the decline was primarily driven a rapid reduction in the effectiveness of their targeted marketing campaigns


To understand specifically what was causing the decline in effectiveness, the first step was to perform a thorough assessment of the situation, focusing on;

  • Customer – what’s changed in the make up the customers continuing to enroll and engage with the marketing and brand

  • Market – what’s changed in the marketplace that could have caused a shift in performance or customer behaviour

  • Process – what’s changed in the processed used? Are models/targeting rules still effective? Any tech or executional changes?

The assessment pointed clear to a couple of different issues – 1) a shift in market dynamics the year prior (driven by a new entrants/competitor) that had stolen share of a specific customer segment; 2) the targeting logic that was set up when the product launched a number of years prior was no longer effective at selecting the customers most likely to be interested.

Solution & Results

A new targeting strategy was devised by firstly sourcing third-party data to significantly enhance our understanding of each customers financial profile and preferences and then secondly building and training a machine learning model to find customers who look like the top existing card holders, using both legacy and the new data acquired.

Implementation of new targeted marketing campaign saw a 4% increase in number of new card enrollments in the first 3 months of implementation

#costefficiency #returns

0 views0 comments