I Knew This Article Was Going To Be About Predictive Modeling

How can marketers know what you’re going to buy before you know what you’re going to buy—let alone what you’ll eat, where you’ll eat it and how much you’ll pay for it? Good marketers get this information not from a crystal ball, but by engaging in database marketing at the highest level. They use Predictive Modeling.

It’s the same reason that the Vegas blackjack dealer and the pit boss share a smile when you step up to their table: They know there’s only a slim chance you’ll be taking home that stack of chips. They note the “classifiers” in your appearance and actions: the clothes, hesitant demeanor and darting eye movements. They see that you’re the opposite of the focused, relaxed gambler in his element.

Marketers identify the classifiers—or attributes—of their ideal customer, including such data as past buying habits, income level and much more. Predictive Modeling is the science of calculating what the model prospect wants and needs based on these classifiers. Predictive Modeling is also used to construct models of likely customer behavior based on particular measures like sales, marketing and customer-retention techniques. In other words, marketers can tell how customers will react to marketing messages, discounts, offers and other actions.

Predictive Modeling comes into play when the market is broken down into segments where customers share common wants and needs. Market segments may include:
a) Groups that have different needs and wants from other groups
b) Groups that have the same wants and needs as other groups
c) Groups that will respond similarly to stimuli in the market

Instead of sending the same marketing message to every potential customer, smart marketers make the most of every dollar by targeting groups most likely to buy and then tailoring their messages accordingly.

Statisticians utilize even more sophisticated elements of Predictive Modeling to further drill down to the exact customer they’re trying to attract. For instance, logistic regression is a sort of “reverse engineering” of Predictive Modeling. And Uplift Modeling is used to predict changes in probability caused by actions for limiting “churn”—natural customer attrition caused by moving, death or a shift in interest.

Predictive Modeling is a very valuable tool that marketers are refining in order to create highly pinpointed marketing campaigns.

But if you use Predictive Modeling, you already knew that.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s