Predictive models allow you to gain insight into your customers by identifying what actions they are likely to take in the future. Below you'll learn about the types of predictive models available and how Listrak enables you to predict the behavior of your customers.
UNDERSTANDING TYPES OF PREDICTIVE BEHAVIORS
Listrak uses multiple types of models to predict the future behavior of your customers.
Affinity: Determines how strong of a relationship exists between the customer and a specific element that motivates their purchase.
Probability: Determines the likelihood that a specific outcome or action will occur. Probability models can output a numerical value or group contacts into buckets.
Lifetime Value: Predicts a customer's potential spend value based on their current lifecycle and spending habits.
PREDICTING PURCHASE BEHAVIOR
Listrak CRM predicts information about a contact's purchases, including what may influence them to purchase (such as an incentive).
COUPON AND DISCOUNT AFFINITY
Coupon Affinity: Groups customers by the chance that they will use a coupon when making a purchase. Passing coupon information in the orders file is required to identify a buyer's coupon affinity.
Discount Affinity: Groups customers by the chance that they will purchase an item using a discount. Passing discount information in the orders file is required to identify a buyer's discount affinity.
COUPON AND DISCOUNT CLASSIFICATIONS
Coupon and discount affinity classifies customers into the following groups:
Coupon/Discount Buyer: 100% of purchases made using a coupon or discount.
Preference: 50 - 100% of purchases made using a coupon or discount.
No Preference: 25 - 50% of purchases made using a coupon or discount.
Full Price Buyer: No purchases made using a coupon or discount.
Non-Coupon Buyer: No purchases made using a coupon.
PRODUCT-LEVEL AFFINITY
Brand, Category, Subcategory Affinity: Groups customers together based on the strength of their relationship with a specific brand, category, or subcategory.
Product affinity classifies customers into five groups: some, medium, medium-high, high, and very high.
CHURN PROPENSITY
Churn Propensity: Groups customers together based on where they are in their lifecycle.
Churn propensity classifies contacts into one of four stages:
Active: Identifies customers who are in the first half of their purchase lifecycle.
At-Risk: Identifies customers who have entered the second half of their purchase lifecycle.
Churned: Identifies customers who have not made a purchase within their expected lifecycle.
Has not purchased: Identifies non-customers who do not have a lifecycle stage because they have not made a purchase.
PREDICTED LIFETIME VALUE (PLV)
Predicted Lifetime Value: The amount a customer is likely to spend with your business within the next 12 months.
PREDICTING ENGAGEMENT
All predictive likelihood models indicate the likelihood that something will occur within the next 14 days.
LIKELIHOOD MODELS
Purchase Likelihood: A calculation indicating the chance a contact will purchase.
Open Likelihood: A calculation indicating the chance a contact will open an email.
Click Likelihood: A calculation indicating the chance a contact will click a link in an email.
Unsubscribe Likelihood: A calculation indicating the chance a contact will unsubscribe from your emails.
LIKELIHOOD CLASSIFICATIONS
Decile: Groups contacts into one of 10 groups. The first decile are the contacts most likely to take an action.
Percentile: Groups contacts into one of 100 groups. The closer a contact falls to 100 the more likely they are to take an action.
Propensity: Groups contacts based on if contact is most, more, less, or least likely to do something compared to the average customer.
Vs. Mean: Groups contacts by comparing their likelihood to take a certain action to the average customer.
Now that you know what predictive models are available in Listrak CRM, you can begin building tiles to gain insights into product-level affinity or the likelihood a contact will engage.