Pilgrim Bank

Pilgrim Bank (A): Customer Profitability – Case Analysis

Problem:

We need to identify if the online customers are indeed better customers and if the adoption of the online channel actually produce better customers in terms of profitability.

Analysis required:

* Establish confidence intervals

* Check if there is significant difference in the profitability between online customers and regular customers

* Study and report how age, income and geographic region affect the profitability of the online customer

Confidence interval: Population size is large and hence we can perform z test to get confidence interval for the average population profit. The confidence interval with 95% confidence:

108.5<profitpopulation<114.50.118<proportiononline<0.125

Online vs Regular Customers:

Null hypothesis is that no significant difference in profitability and Alternate hypothesis would be that significant difference exists

H0:profitonline-profitoffline=0H1:profitonline-profitoffline≠0

t test analysis of comparison between online and regular customers.

t-Test: Two-Sample Assuming Unequal Variances |

  | Online customers | Offline customers |

Mean | 110.79 | 116.67 |

Variance | 73604.22 | 80465.63 |

Observations | 27780.00 | 3854.00 |

Hypothesized Mean Difference | 0.00 | |

Df | 4882.00 | |

t Stat | -1.21 | |

P(T<=t) two-tail | 0.23 | |

t Critical two-tail | 1.96 |   |

t value is -1.21 which lies in the acceptance region of -1.96 to +1.96.Therefore null hypothesis is accepted and we can conclude that there exists no significant difference in profitability of online and regular customers.

Therefore there is no difference between profits of online customers and offline customers.

Effect of the various independent variables: Profitability is the dependent variable and online usage, age, income, tenure and geographic area are independent variables. First regression run showed that p-value for geographic region was as high...