Wouldn’t it be great to be able to predict when a customer is likely to make another purchase from you? Or to know which of your other products are likely to interest them, or be able to make an educated guess about some other aspect of their behavior? Good news: you can! How? By using Predictive Modeling.
What is Predictive Modeling?
Predictive Modeling is a way to use all of your existing data about your customers and prospects – and their behavior in different marketing/sales situations – to create a statistical model of future behavior. Information about a particular type or segment of people can be generalized to all others of that type or segment, and can be personalized to a particular person, or a combination of both.
Amazon.com is a master of Predictive Modeling. When you search for something on Amazon they immediately give you suggestions for other products that might interest you as well. These suggestions are based both on what they know about your search and purchase history, and what they know about others who have searched for or purchased similar items and/or have other characteristics in common with you.
Is Predictive Modeling Always Right?
Like all statistical models, Predictive Modeling sometimes leads marketers awry. For many years I lived in an area in which nearly all of my neighbors had kids. Based on Predictive Modeling many companies assumed that I was a parent, too. I wasn’t. Which means that all of the offers for kids’ meals and math tutoring that arrived in my mailbox went straight to the recycling bin.
How Can You Use Predictive Modeling?
A popular use for Predictive Modeling is for figuring out when a particular customer is likely to make their next purchase, so that you can determine the best possible time to send them a personalized offer about this purchase. Predictive Modeling is also excellent for cross-selling or up-selling (as in the Amazon example).
My husband and I recently purchased a new home and instantly we started to receive catalogs and coupons for household items from local retailers. In the home improvement industry, for example, Predictive Modeling has been used to determine that people who buy a new home are likely to purchase household items from stores like Home Depot within the first 3 months of their purchase. This type of information is invaluable for shaping and timing marketing campaigns.
