# 5 bucket gains chart – look B u s i n e s s F i n a n c e

5 bucket gains chart – look B u s i n e s s F i n a n c e

•That’s Entertainment is a club marketer of videos. They are testing a new music club concept. A 25,000 sample of names from the That’s Entertainment database was test mailed for this brand new music club concept. For those names that joined the new club, they received 10 free CDs and agreed to purchase 2 more CD’s over the next 12 months.

•The test had a response rate of 40% for the initial offer of 10 free CD’s. All customer data was saved point-in-time of the promotion for future analysis purposes.

•That’s Entertainment has decided to roll-out with the new music club concept. They do not wish to promote all names on their customer database. As such, they have requested the build of a response model to help them select the names most likely to join the club.

•Using the frozen file, you will build a multiple regression response model predicting who is most likely to join the new music club. You will use Excel for this exercise and base the analysis on a sub sample of 150 names randomly drawn from the 25,000 sample.

•Using Excel(Go to HW folder for That’s Entertainment excel sheet) and perform the following tasks:

–1. Run a multiple regression model using all three variables simultaneously (TSLO, DOLL_CR, and NM_ORD) as your predictors and using the order indicator (ORDER) as the dependent variable.

–2. Examine the output. Do you see any problems with the coefficients that may be due to multicollinearity? If so, run a correlation analysis to confirm. What do you notice?

–3. If there is a problem with one of the variables being correlated with another, determine which variable to delete and rerun your model. Explain how you determined which variable to delete.

–4. Once all issues of multicollinearity are taken care of, examine the p-values associated with your predictor coefficients and comment?

–5. What is your final model?

–6. Now do the same thing with the cars data set in SAS (from SASHelp library) with at least four variables as predictors and one dependent variable. Interpret results.

Compute a 5 Bucket gains chart

Look into Stepwise/Backwards/Forward Regression and apply to this project.