Customer acquisition – SAS output

Notes

1. A random guess model would do the following. First, it would determine which bucket acquire or not acquire has more prospects in it. In this case 292 prospects were acquired versus 208 who were not. Then it would predict that all prospects would be acquired and it would be accurate 58.4% of the time (292/500).

2. There are many other distributions which can be used including the exponential, gamma, normal, and log-normal among others. It is often worthwhile to try several distributions to see if the model fit and prediction improve.