A predictive model is a system that learns from past events to anticipate what is likely to happen. In marketing, it analyzes the history of thousands of customers, identifies patterns, and applies them to each individual to estimate, for example, whether they will make another purchase, when they will do so, or how much value they may bring in the future.

Why does it matter?

Most marketing decisions look back: who bought, how much, and when. A predictive model lets you look ahead and take action before things happen.

• Prevent: Identify who you're about to lose while you can still hold on to them.
• Anticipate: Reach out to each customer around the time they usually make another purchase.
• Prioritize: Invest more effort in the customers with the greatest future value.

A prediction isn't a certainty—it's a probability. Its value lies in helping you set priorities based on more information than just intuition.

Story mode

An online wine store sends its monthly promotion to its entire customer base. Using predictive models, it changes its approach: it shows customers with a high propensity to buy new arrivals without a discount, and offers those with a high probability of churn an incentive to return. For example, Diego, who used to buy every month but hasn't done so in three months, receives a selection of his favorite wines.

How to Use It with Data4Sales

Data4Sales includes predictive models based on each customer's 360° profile:

• Churn probability.
• Probability of a repeat purchase (propensity).
• Days until the next purchase.
• Predictive LTV, which estimates the future value of customers with over 90% accuracy.

You can use these predictions as filters to build audiences and activate them in workflows and campaigns via WhatsApp, email, or SMS.

Related terms

• Churn Probability
• Willingness to Buy
• Days until the next purchase
• LTV