Demographics, Models and Your Bank

By Tony Rizzo

Providing the right products, services, and messages to the right customers at the right time is at the core of marketing. And here’s where bank marketers have an edge: the immense stockpile of relevant data that banks generate makes it that much easier to target the right customers at the right time. The question is how best to tap this resource.

Start with demographics and predictive models.

By definition, marketing segmentation is an attempt to cluster groups of customers based on like qualities, domicile, psychological personas, and/or behavioral traits. Demographics tend to be the go-to data resource for most small to medium financial institutions. This is not necessarily a bad thing, as demographics allow institutions to build accurate profiles from which an institution can build the product offering, marketing messages, and overall brand.

Primary demographic elements include income, presence of children, rural versus urban designation, and homeowner versus renter status. With these simple elements an institution can learn enough about its customer base to effectively segment when combined with internal data gathered from the customer relationship management (CRM) tool or marketing customer information file (MCIF).

Next level demographics are those that seek to define predictive habits. Data elements of this type can include propensities across a wide range of retail financial services. These predictive models also provide insight in terms of numerical ranges regarding how interested consumers are in purchasing a particular product and when they will do so. For example, to quickly target customers interested in a home equity line of credit, a numerically scaled and attribute-weighted predictive model can account for a range of characteristics (income, home ownership, estimated home equity, length of residence, presence of children, and repayment ability) without having to individually correlate discrete data points on a repetitive basis. This allows for speed of execution.

How does it work?                                                          

The effective use of demographics and predictive models is a two-step process.

  1. Identify the bank’s most valuable customers based on internal data (profit, number of products, tenure, domicile, and transaction history) and external data (life stage code, credit score, income producing assets outside the bank, and home value). Once that’s complete, look for trends in the appended demographic or predictive model data. After this analysis, the institution should understand who its most valuable segment is from internal and external perspective.
  2. Use the analysis from step one to discover look-alike customers from a demographic standpoint. For example, if an institution finds its most valuable households to have annual incomes of more than $75,000, gather customers outside of the primary segment who meet this demographic criterion. This analysis, often referred to as “filtering,” should produce a final segment of potential customers who are similar to the bank’s best customers but require additional communication in order to increase their activity via product purchase. One’s CRM or MCIF tools will also produce cross-tabulation reports allowing for very sophisticated statistical analysis without the pain of having to understand the mechanics of the calculations.

Consider the retail lending potential.

For those institutions intensely focused on the acquisition of retail and real estate loans, the best way to acquire those loans is from active and engaged customers. By working with an outside vendor who has aggregate credit report data, an institution will know the number—and balances—of auto loans, credit cards, home-equity loans, installment loans, and mortgages a current customer has at other financial institutions. Armed with this knowledge, the obvious deliverable would include a series of communications designed to migrate those existing loans to the bank’s balance sheet. The table below shows a summary of this type of analysis executed for a billion-dollar in assets based in Pennsylvania. It demonstrates that for every one loan owned by the institution, 12 were held outside the institution. Additionally for every $1 the institution held on its books, $19 in potential loans were available at other financial institutions. The importance of this analysis demonstrates the large amount of potential business a bank has simply by focusing on its current customer base.


(Click on table to enlarge)

Consider the commercial lending and deposit potential.

Commercial business is one of the most difficult relationships to acquire. The goal is to establish a mutually profitable, long-lasting business relationship. Typically this relationship is conceptualized as a construct composed of commitment, satisfaction and trust. From a direct marketing standpoint, these paradigms are difficult to effectively convey with the printed word in such a fashion as to generate significant returns on the marketing investment. Therefore, it makes sense to seek out those business owners who are current retail customers of the financial institution.

Going a step further, it makes sense to seek out those business owners who are active, involved, and highly profitable retail customers of the financial institution. By seeking out data of this nature, and acting upon it, the financial institution is able to move the sales process along much faster as the end consumer (the current retail customer) already has intimate knowledge of the financial institution but perhaps is unaware of the commercial services offered by the bank. By appending this type of data to the institution’s CRM or MCIF analytical tool, the institution will have a detailed understanding of the market potential available to it as well as the intimate knowledge of which customers they should speak to immediately.

Appended data elements of this type not only include the usual suspects of NAICS (North American Industry Classification System) codes, revenue, number of employees, and length of business, but also propensity models across a wide range of commercial products. Using the same institution that was analyzed for retail loans, it was discovered that an additional $500 million in both deposits and loans were available from retail customers who own a business where there was no current commercial relationship. In other words, as you can see on the retail loan analysis, significant growth is available by simply looking at the current customer’s life outside of the institution and taking action.

Tony Rizzo is General Manager of Creative Services at Marquis, a Plano, Tex. data analytics company specializing in the financial services industry. A 26-year veteran of financial marketing, Tony has been published in every major financial industry publication as well as the Wall Street Journal and USA Today, and is a frequent national speaker on leveraging data to increase retention and profitability from a marketing perspective. Email: [email protected].

Hear more from Tony when he presents at the ABA Bank Marketing Conference, September 25-27, 2016.  We look forward to seeing you in Nashville!