ANALYHG CHAUENGE #1: RFIVI Model Enhancement via Binning m /\\ H\" \"if Your manager at Bookbinders is impressed with the improved profits that can be generated by using a simple RFIVI model compared to mailing all of her customers. She is wondering, however, whether the RFIVI model {independent ntiles] can be improved by adding one additional binary variable. We looked at an example in the video for this challenge in terms of binning the child variable {2 to 3 = 1; else 0]. Your task is to create, via binning, a different binary variable and add it to the HFIVI model with the goal of enhancing the model's ability to nd buyers and thus enhance protability. You will need to choose which variable you wish to bin [i.e., dichotomize] and then bin it into a binary format. If the variable you are binning is, for example, number of children's books purchased in the past, then you would: Transform... Recode into Different Variable... Input Variable=child; Output Variable = V; [click the Change button]; Click on Old and New Values....and enter the logic. For example, if you wish to create a new variable that equals 0 ifsomeone has bought 2 or 3 children's books in the past, and 1 if they have bought any other number of children's books, then you would add this logic {Old Value, Range: 2 to 3, New Value :1; click Add; l0|d Value, mother values, New Value = 0; click Add]. You need to provide rationale for Eur variable and for the split point u chose [why should 2 and 3 children's books be one category, and all other values another category, for example]. You therefore need to start by creating R, F, and Min BBB.sav, as we have done in the past. Then create your new binned variable. Then create composite codes that will contain your new binned variable as the fourth element of the code, RFMV {let V represent your new binned variable, and code it as 0 or 1].l Then, run case summaries, analyzing buyer (N, sum, mean] by your RFMV code, transfer the results to an excel le, sort the rows [excluding total row] by response rate, and calculate the protability of mailing all RFMV segments at or above breakeven response rate [.033 at Bookbinders]. 1 You will need to multiply it by 1000, F by 100, M by 10, and then add V, since there are now 4 digits in your RFMV oode. Keep in mind that at Bookbinder's, it costs $.5D to print up and mail a catalog to a customer, and the firm earns 56 in contribution per buyer. Be sure to answer all ve questions below. A template for your answers is attached to this document on page 3. Report: 1] 2] 3] Which variable did you decide to bin, what was the split point, and why? Here is where you provide rationale for your variable and for the split point you chose. Using your enhanced RFM'v' model, how many customers will you mail? How many buyers will you obtain? How much money will you make? By how many dollars does your enhanced FlFlv'l'vIr model beat the baseline RFM {independent ntiles] approach? [Note: The independent ntiles EFM model generates 51,9134 in gross prot, for 513556 RUMI, in BBB.sav.] Insert a screen shot of that part of mexcel spreadsheet at and above breakeven, with your prot calculations {for your enhanced model] as support; or attach your excel spreadsheet with this information. Write a brief executive summary (1 or 2 sentences] for your manager. If your enhanced model's profits are higher, address whether you feel the added complexity of your model is "worth it." {Remember that your model would be applied to millions of customers who were not included in this 5D,DDD customer test mailing.) ANALYTICS CHALLENGE #1: RFM Model Enhancement via Binning Your Answers 1) Model description a. I decided to bin the variable. I chose this variable because: b. The split point I chose was 0 = I chose this split point because: 2) Using my enhanced RFMV model: a. I will mail customers. b. I will obtain buyers. c. Gross profits will be: $ 3) My enhanced RFM model beats the baseline RFM (independent n-tiles) approach by $ 4) Below I attach a screenshot of my spreadsheet at and above breakeven, with profit calculations (for the enhanced model) as support:5) Executive Summary