Customer retention – SAS output

Notes

1. A random guess model would do the following. First, it would determine which bucket acquire or not acquire has more customers in it. In this case 2678 customers purchased versus 2822 who did not. Then it would predict that all customers would not purchase and it would be accurate 51.3% of the time (2822/5500).

2. A random guess model would do the following. First, it would determine which bucket cross-buy or not cross-buy has more instances in it. In this case 1361 cross-buy occasions happened versus 1317 that did not. Then it would predict that all customers would cross-buy and it would be accurate 50.8% of the time (1361/2678).