What Risk Probability Means

Every active customer has some probability of defecting, however small, measured by their Risk Probability.  This is the probability, between zero and 100%, that a customer will not make a purchase in the time between the most recent analysis date and the date when their recency exceeds the length of the inactivity period (in other words, when they become inactive and are considered defectors).  In a group of 100 customers all having risk probability of 90%, we expect approximately 90 of them to actually defect and approximately 10 of them to make a purchase and remain active. A high risk probability does not mean the customer will defect, but it does mean that the customer is more likely to defect than a customer with a lower risk score.

IMPORTANT: For better granularity, the Risk Probability scale runs from 0 to 1000 with three significant figures, rather than from 0 to 100. For example, if you want the lower bound to be 75%, set the left end of the slider at 750.

The goal of the Risk Analyzer is to identify any potential defectors as soon as possible, so you can market to them to keep them active.  But labeling every customer as a potential defector is counter productive, since there may be little need for discounts for some customers.  Thus it is important to have an accurate probability calculation, and also important to set the threshold such that the right number of customers is selected.  The default setting of the Longbow Risk Probability metric will select approximately the same number of customers as go inactive each period.  See the Risk Analyzer Primer for a more detailed discussion of this setting and the other metrics.