Challenge
A medical device mailer was having great success targeting audiences within Alliant’s consumer database, using a custom conversion model. The client touted the audiences as their “best performing” prospects, driving a consistently high ROI. The problem was: they wanted more. And digging deeper into the model meant weakening the performance they loved.
Solution
Alliant needed a strategy to increase the number of top-ranking prospects capable of maintaining its high KPI benchmarks. To do this, Alliant data scientists decided to use the power of three predictive models instead of one: the original conversion model plus a new clone and decision tree algorithm. With Alliant’s deep sources of transactional and product interest data, the team was confident the solution would find more promotable prospects.
Results
With the new approach, Alliant grew the prospect pool by 50% and — even more remarkably — maintained its abnormally high ROI and placement as the client’s #1 performing data source. It’s clear that by combining a trio of modeling techniques Alliant is able to deliver more qualified prospects, above and beyond the scope of an already strong model.
