Predictive Analysis involves projecting the future behaviour of customers based on past history.  It involves correlating behavioural pattern with groups of customers called customer segments.

Each customer is grouped into one or more customer segments.  Although customer segments are often specific to an industry, some segments are common to most industries:

1. physical location segments (eg. mid west, south, south west)
2. gender (male, female)
3. age segments (e.g. under 15, 15-20, 20-30, 30+)
4. education segments (e.g. high school, college, university)
5. income segments (e.g. <$20,000, $20,000-$40,000, 40,000)
6. transaction value segments (e.g. <1000, <$500, $0)
7. account balance segments (e.g. $0, >$500, >$1000)
8. channel usage segments (e.g. phone, web, in-person)

By analyzing the historical behavior of groups of customers, it is possible to predict the future behavior of customers.  This helps businesses answer they following types of questions:

1. Which customers are likely to default on loans?
2. Which customers are likely to be highly profitable to the business?
3. Which customers are motivated by sales?
4. Which customers are likely to be lost to competitors?
5. Which customers are likely to purchase other products or services?