Data
Data is split 70/30
Total of 10127 observations
We are predicting how many customers leave and stay
##
##
## Cell Contents
## |-------------------------|
## | N |
## | Chi-square contribution |
## | N / Row Total |
## | N / Col Total |
## | N / Table Total |
## |-------------------------|
##
##
## Total Observations in Table: 3039
##
##
## | val$Attrition_Flag
## val$predicted_rf | Attrited Customer | Existing Customer | Row Total |
## ------------------|-------------------|-------------------|-------------------|
## Attrited Customer | 388 | 20 | 408 |
## | 1653.315 | 305.525 | |
## | 0.951 | 0.049 | 0.134 |
## | 0.819 | 0.008 | |
## | 0.128 | 0.007 | |
## ------------------|-------------------|-------------------|-------------------|
## Existing Customer | 86 | 2545 | 2631 |
## | 256.386 | 47.379 | |
## | 0.033 | 0.967 | 0.866 |
## | 0.181 | 0.992 | |
## | 0.028 | 0.837 | |
## ------------------|-------------------|-------------------|-------------------|
## Column Total | 474 | 2565 | 3039 |
## | 0.156 | 0.844 | |
## ------------------|-------------------|-------------------|-------------------|
##
##
Financials
We predicted 473 were going to leave
The cost of the program would be 473 x 100 = $47,300
Saving 30% on 415 would be 124.5 with a value of $2300 x 124.5 =
$286,350
Net Profit of this would be $286,350 - $47,300 = $239,050
Profit per customer is $239,050 / 3074 = $77.77