Branch Power at Fleet
For the last few years, banks have been exploring the potential for customer relationship marketing with very productive results. The process goes this way:
This is the story of how Fleet Bank went through these steps, what they learned, and how they are going about providing their branch personnel with the information about their customers and the training necessary to improve customer profitability. The process was directed by Randall Grossman, Senior Vice President and Director of Customer Data Management and Analysis of the Fleet Financial Group,
Fleet had developed a bank-wide MCIF in the early 1990s, maintained externally on a Harte-Hanks database. In January 1996, they took their first measure of customer value by simply adding up all non-mortgage deposit and loan balances for each customer. By the end of 1996, they had created the software necessary to determine the Net Income after deduction of the Cost of Capital (NIACC) for all retail customers. For a typical customer, it looked something like this:
In this annual calculation for a typical customer, her net profitability is $63, due principally to the fact that she invested in mutual funds. Without this investment, she would have represented a loss of $166 to the bank.
By the end of 1997, Fleet had extended this system to its commercial customers. They were using industry benchmark costs for computing profitability. By the end of 1998, the bank had created an in-house data warehouse which enabled them to keep all of this data current, and to use actual Fleet activity based costs. The system looked something like this:
At their PC workstations using Windows NT and 95, power users, marketing and business analysts can access the data on each customer, can do analytical work and modeling, and develop marketing initiatives. The analysis server, for example, provides for statistical analysis, neural networks, ad hoc query and analysis, and geodemographic analysis. The data mart provides summarized, pre-formatted data for promotion design, tracking and analysis, and enables the users to do point & click, drill down analysis. The management reporting server provides on-demand parameterized reports.
Preliminary Lessons Learned
The new system enabled bank management, for the first time, to really understand their customer profitability, and to do something about it. The data showed that:
Even though half of all customers were considered unprofitable, Grossman realized that Fleet could not simply walk away from half of their customer base. Further analysis showed Grossman that:
Faced with these sobering facts, the marketing staff decided to figure out ways to use the database that they had created to turn the situation around. The key to Grossmans strategy was to develop three measures of customer value:
Lifetime profitability is the net present value of the expected future stream of net income after the cost of capital, discounted at the corporate hurdle rate. It is calculated based on the current products that the customer is now using, including planned re-pricing. As calculated by Fleet, customer profitability differs from organizational profitability for several reasons:
Given these qualifications, lifetime customer profitability at Fleet is calculated for each bank customer and stored in the customers database record each month. Customers can then be ranked and segmented. It is possible to pick out the top customers, those just below the top, the average customers and the unprofitable customers. They can be flagged in the customer record so that marketers and branch personnel can recognize their value to the bank, and develop appropriate strategies and tactics.
Potential Profitability carries the analysis one step further. A typical customer has a limited number of bank products. There are usually many other bank products that the customer could be using. The probability of a given customer purchasing an additional product can be determined using CHAID analysis. For example, if a customer owns a home with a mortgage of $W, has a checking account with an average balance of $X , a savings account with an average balance of $Y, and a monthly credit card usage rate of $Z, an age of 44 years, two children in college, CHAID is used to predict the likelihood of him purchasing:
CHAID can also predict the average balance that he will maintain on each of the possible additional products. Logistic regressions are then used to determine the expected Net Income after the Cost of Capital (NIACC) that Fleet will realize from the possible sale of each of the products to the customer. In each case, an estimate is made of the promotional expense involved in getting him to purchase the product. The potential profitability, then is calculated for each product as the:
Probability of purchase x Expected NIACC from usage Promotional expense
The software then adds up each of the products for this customer with a positive NIACC to get the potential profit.
Potential Customer Value is then determined for each customer by adding together the lifetime profitability (with current products) and the potential profitability (from possible new products). This value is stored in every customers database record and used to select the most likely candidates for promotion for each product in direct mail promotions. It is also used to suggest the next best product when branch personnel are talking to the customer, or customer service has them on the line.
Mobilizing Branch Personnel
Knowing potential customer value and keeping it in a database is useless unless the data can be put to work by customer contact personnel. The central marketing staff can use the data:
Beyond that, however, the bank seeks to mobilize its branch personnel to use the new customer data.
The bank has 1,200 branches, many of which have more than one officer that is interested and skilled at identifying targets for bank programs, and determining the appropriate tactics for each particular case. Here is where branch empowerment becomes important. Too many marketing staffs assume that they know what is best. Furthermore, most direct marketing professionals like to speak in terms of mailings of hundreds of thousands. They use statistical programs on computers to determine what should be done. Branch personnel, on the other hand, who know their customers, and see them once a week or more often, can identify those customers who they spot as being obvious candidates for particular products. They can dream up creative ways of suggesting these products to their customers. They are comfortable with programs that identify the ten or twenty best customers for action in a given week. By mobilizing their imaginations, and entrepreneurial skills, the bank can put their calculations of potential customer value to work.
Banks still have a long way to go in training their staffs to use potential lifetime profitability in their work. But, at Fleet, the data are now available. They are poised for a big boost in profitability by knowing their customers, and training and empowering their customer contact personnel to use that knowledge.