Blog

Executive SummaryIn this report, we analyze the factors effectiveness of marketing activities in retainingthe customers of a retail bank. The customer lifetime value (CLV) was used tomeasured the profitability of customers to the bank. Several data points werecollected on each customer, to analyze their impact better. These include loyalty,value, satisfaction, gender education levels, income, savings & loans taken at thebank, duration of relationship, age of customer.Through our analysis, we find that loans taken at the bank, income of the customerare the best predictors of customer lifetime value. The other factors have verymarginal to none, effect on customer lifetime value. Thus, to increase the CLV ofcustomers, the retail bank has to focus on increasing the customers with higherincomes, and ensure that the customer take higher percentage of loans from thebank.IntroductionThe objective of this project is to develop a reliable regression model to test theeffectiveness of marketing activities in retaining the customers. The dependentvariable in the analysis is the customer lifetime value (CLV). Customer lifetime valueis the metric which measures the profitability of each customer to the retail bank.Higher the customer lifetime value of a customer, more profitable the bank would be,if they kept the customer with them as a client. Banks should aim their marketingcampaigns at increasing the number of clients who have higher customer lifetimevalue, so that profits of the bank can increase. Higher customer lifetime value (CLV)clients are more prized by the firms as they increase the profits for the firm.There are several independent variables to be considered for the regression model.This includes loyalty (measured on a 7 point scale), value, satisfaction (bothmeasured on a 10 point scale), gender, education levels of the customer, income(bracketed into 5 categories), savings held at the bank, loans held at the bank, theduration of the relationship between the customer and the bank, age of thecustomer.Descriptive Stats of the variablesGender of the respondentsThe following table shows the frequencies of different genders of the customers areregistered at the retail bank.There are 297 male customers in the data set and 433 female customers in the dataset. While there was one datapoint where the gender was marked as “0”, we imputedthe value to be “2”, as “2” is the mode of the distribution. In the codebook for the dataset, it is mentioned that “2” denotes female customers. Thus, we have 434 femalecustomers and 297 male customers.The majority of customers in the data set are female. In terms of percentages,59.37% of customers in the data set are female while the remaining 40.63% ofcustomers are male. The gender of the customers skews female in the data set.BranchThe following table shows the frequencies of different branches at which thecustomers are registered at the retail bank.The data set contains data from 6 branches of the retail bank. The highest number ofdata points originate from the 4th branch, with 183 customers in the data set whilethe 6th branch is responsible for the least number of data points, with 29 customersin the data set. The number of customers at the first five retail branches are similarwhile the 6th branch has lower number of customers in the data set.Customers’ IncomeThe following table shows the frequencies of different income brackets of thecustomers at the retail bank.The most common income of the customers at the retail bank is income bracket 3,which is folks making £15,001 – £25,000. There are 73 customers in the data setwho make more than £35,000. There are 65 customers who make between £0 and£5,000.The bank could possibly target more customers who are in the higher incomebrackets, as they are likely to provide higher customer lifetime value (CLV), as theyare more likely to borrow higher amounts of money and repay them on time as theyhave more disposable income with them.Education levels of customerThe following table shows the frequencies of different education levels of thecustomers at the retail bank.The most common education level at the bank, is category “3”, which indicates thatthere are 272 people with BSc degrees, followed by category “4”, which indicatesthat there are 223 folks with MSc degrees, followed by category “2”, which indicatesthat there are 183 folks with College degrees. There are 18 customers in the data setwho possess PhD degrees.The overwhelming majority of the customers at the bank have attained educationalqualifications of college degree or higher.Age of customersThe following table shows the summary statistics of ages of the customers at theretail bank.The youngest customer at the retail bank is 14 years while the oldest customer is 79years old. The median age of customer is 45.14 years, while the mean is close to themedian, with an average customer age of 45.21 years. The 25th percentile of age is36 years of age, while the 75th percentile of age of customers at the bank is 54years. The bank mostly lends to middle aged customers, while there are customerson each of the age spectrum. The bank could focus more on richer retirees who aremore likely to have better financial positions, and deliver better customer lifetimevalue (CLV) and increase the bank’s profits.Duration of relationship of bank with its customersThe following table shows the summary statistics of duration of relationship of theretail bank with its customersThe shortest duration of customer relationship at the retail bank is 1 month while thelongest duration of customer relationship is 240 months or 12 years. The medianduration of customer relationship is 24 months or 2 years, while the mean duration ofcustomer relationship is close to the median, with an average duration of customerrelationship

of 31.48 months. The 25th percentile of duration of customer relationshipis 12 months or 1 year, while the 75th percentile of duration of customer relationshipat the bank is 36 months or 3 years. The overwhelming majority of customerrelationships at the bank have been less than 3 years old. There are somecustomers whom the bank has known for a long time, but the majority of thecustomer relationships are new, and indicates that there is potential to improve therelationships by establishing longer relationships.Longer relationships, coupled with good customer service will translate to morecustomers moving their savings account to the bank, while taking loans from thebank. These moves will likely increase the customer lifetime value (CLV) metric ofthe customer and increase the bank’s profitability in the long-run while ensuringstability of revenues due to the long-term relationships established.% of savings held at bankThe following table shows the summary statistics of percentage of savings ofcustomers held at the bank.The average percentage of savings held at the bank by a customer is 64.73%, whilethe median is 75%. The 75th percentile is 100% and the 25th percentile is 30%. Thebank needs to deepen its relationship with the customers, so that they transfer thesavings to accounts opened at the bank. As bank gains access to more deposits andsavings, its current and savings account ratio increases, as these accounts enablethe banks to access funds at very low cost of funds.As more customers transfer their savings to accounts held at the bank, there isincreasing chances of cross-selling happening, where the bank pitches additionalservices such as loans, insurance, fund transfer, payments and other financialservices to these customers, which would increase the revenues and profits for thebank. Through these extensive services, the customer is more likely to turn to thisparticular bank than any other bank for their financial services needs, and createslasting long-term relationships between the customers and the banks.% of loans held at bankThe following table shows the summary statistics of percentage of loans ofcustomers held at the bank.The average percentage of loans held at the bank by a customer is 52.71%, whilethe median is 52.90%. The 75th percentile is 52.90% and the 25th percentile is 50%.The stats suggest the most customers in the data set hold close to 53% of theirloans at the bank. This is a healthy measure as half of the customers loans’ aretaken at the bank. There are limitations to increasing the percentage of loans held atthe bank beyond certain thresholds, as it leads to unnecessary increase inconcentration risk for the bank. Diversified loan portfolios of a customer decrease thechances of defaults for the bank on a customer, and provides reasonable protectionto the bank.Customer Lifetime Value (CLV)The customer lifetime value varies from a minimum of £1,181 to a maximum of£27,995. The median customer lifetime value is £5,087 while the mean customerlifetime value is £6,704, significantly greater than the median value of customerlifetime value. The 25th percentile of customer lifetime value is £3,805, while the 75thpercentile of customer lifetime value is £7,947. It is likely that higher value ofcustomer lifetime values are skewing the mean value, which results in greater valueof mean customer lifetime value than the median customer lifetime value.The boxplot of the customer lifetime value (CLV) variable is shown in the figureabove. As indicated by the summary stats, the higher value of mean customerlifetime value compared to median customer lifetime value is explained by thepresence of the significant number of data points in the right tail of the distribution,closer to the maximum values, which swings the mean customer lifetime value (CLV)higher than the median value of customer lifetime value.We would like to explore whether customer lifetime value varied by the gender of thecustomerThe customer lifetime value (CLV) values doesn’t seem to be clearly differentiated bygender of the customer, as the median values of customer lifetime value are close byfor both the genders.The average customer lifetime value for male customers is £6,883.116, while forfemale customers at the bank, the customer lifetime value is £6,591.719.We have plotted the boxplot of customer lifetime value by the income groups, shownin the plot above. There is clear differentiation of customer lifetime value between thedifferent income groups. The higher median customer lifetime value is witnessed forincome groups 4 & 5 relative to other income groups. The income group 1 has theleast average customer lifetime value, while income group 5 has the highestcustomer lifetime value.The different mean customer lifetime value levels are seen in the income groups,shown above. The income group 5 (> £35,000) has mean customer lifetime value of£8804.296, followed by income group 4 (£25001 – £35000) with mean customerlifetime value of £7710.275.Reliability Studies using Cronbach’s AlphaCronbach’s alpha was calculated for the Loyalty scales. The metric is equal to 0.71,which is greater than 0.7 & indicates the scale is reasonably reliable.The Cronbach’s alpha was calculated for scales measuring Satisfaction. TheCronbach’s alpha is 0.76 > 0.7, which indicates good reliability for the scales used.Data Analysis using Regression modelsWe anticipate that the bank has higher customer lifetime value (CLV) with customerswho have taken large loans from the bank, those have longer relationship with thebank and those who belong to higher income brackets. We take these as theindependent variables in the first regression model.The adjusted R-squared value is 0.1101 & the F-statistic is 23.74, degrees offreedom = (4, 731) and the associated p-value is 2.2e-16