Original article was published by Frederik Bussler on Artificial Intelligence on Medium
How to Optimize Customer Conversions With AutoML
Data-driven sales on a banking dataset.
81percent of sales leads aren’t converted¹. Without targeting the right leads and optimizing conversions, getting more leads is like pumping water into a leaky bucket.
In another article, I explored how to use machine learning to optimize enterprise sales conversions. Here, we’ll explore how to optimize customer conversions from a banking dataset using the predictive analytics tool Apteo.
We’re given customer attributes like age, job type, duration with the company, account balance, and so on, which we can use to predict
subscribe, or whether the customer subscribes to a term deposit.
This is an example of a binary classification machine learning problem, since
subscribe can either be “yes” or “no.”
Loyal Customers Convert More
The longer a customer has been with the company, the higher their odds of converting to a new product or service. The most important attribute in explaining
subscribed is the
duration column, or simply how many days the customer has been with the company.
This holds true for any business relationship. Employees with a longer tenure have a lower chance of attrition and customers with a longer tenure have a lower chance of churn. Below, we can see that the average customer who said “yes” has been with the company for much longer than customers who said “no.”