Pilgrim Case

A. Is there a difference in profitability across regular vs. online customers?

For both years, 1999 and 2000, the Online customers had the greatest average profit; also there was an increase in the number of Online customers. In fact, the total Profit from Online customers almost doubled since 1999. On the surface, banking online does seem to drive profitability, or at least, should be reason to explore the relationship further.

 B. How should we deal with the missing data?

Regression substitution, Mean substitution and stochastic regression can be used to deal with missing data.

Regression substitution: Using linear regression to predict what the missing score should be on the basis of other variables that are present? We use existing variables to make a prediction, and then substitute that predicted value as if it were an actual obtained value. Stochastic regression imputation, This approach adds a randomly sampled residual term from the normal (or other) distribution to each the imputed value.

Mean substitution: procedure that certainly be relegated to the past was the idea of substituting a mean for the missing data. In the first place it adds no new information. The overall mean, with or without replacing my missing data, will be the same.

C. Using a regression model, is there a difference in profitability across the two customer types and if so, how much are the more profitable customers worth to the bank?

In order to explain the relationship between a customer following “regression” type relationship can be used:

Customer Profitability = f (Online, Demographic variables)

The idea of building this regression model is for the purpose of explanation more than

Prediction. The drivers of profitability for the year 1999 with a very simple Linear Regression Model using all demographic variables as inputs, including whether the customer was an Online user and if they used Pilgrim’s online BillPay service. What we get is the following equation:

 ...