Dataset used: "HW-LOGISTIC-LOYALTY.SAV"

Data Description: A Fortune 500 service company has collected data to build a predictive model of customer loyalty to answer the question: what factors are associated with high versus loe customer loyalty? The managerial decision is to invest in levers of organizational action that are likely to win customer loyalty.

Customer loyalty is based on archival records of past behaviors and is classified into high or low loyalty (BLOY).

Potential predictors include customers' satisfaction with company's services, perceived value in ongoing relationship with the company, trust in the company's business practices and the agents that interact with them, perceived reputation of the company in the industry, and evaluation of the price-value comparison of the company's services (variable names are underlined and bolded).

The data also includes customers' demongraph variables including age, sex, and education. All customers in the data are US-based.

Purpose: To demonstrate skills for data mining, and logistic regression modeling for categorical outcomes. Also to interpret results from a managerial point of view.

Specific SPSS Procedures Used: Frequencies, Cross-tabs, Recode, Graph, Logistic, Classification Accuracy, ROC

Company's Hypothesis: The company's hypothesis is that customer loyalty will be predicted by satisfaction, trust and value, but that the predive effects will vary for males and females after controlling for education and age.

Test, interpret and evaluate the company's hypothesis by building a robust logistic model for empirically testing it. Be sure to examine the basic frequencies and distributions of the data to ensure that it is free of errors.

Improve the predictabiity of the customer loyalty model in 1 above, by including other predictors and interactions terms that were not included in the company's hypothesis but are empirically effective and useful.

Evaluate the classification accuracy of the company's hypothesized model relative to the model you developed in 2 above.

Based on your analysis, draw critical insights for managerial action.