Editor’s note: Crestar Financial Corp.(formerly United Virginia Bankshares, Inc.) is the holding company for Crestar Bank of Virginia, Crestar Bank N.A. of Washington, D.C. and Crestar Bank MD of Maryland (formerly United Virginia Bank, NS&T Bank, N.A. and Bank of Bethesda, respectively). Headquartered in Richmond, Va., Crestar serves customers through a network of 234 banking offices and 145 automated teller machines. It offers a broad range of banking services, including various commercial and consumer loan and deposit instruments, trust services, bank credit cards and international banking. Crestar's subsidiary, Crestar Insurance Agency, Inc., offers personal auto and homeowners insurance as well as a variety of annuities and life insurance products. Discount brokerage services are offered by Crestar's subsidiary, Crestar Securities Corp. First and second mortgage loan origination and servicing and Capitoline Investment Services Inc., respectively. The non-banking subsidiaries provide services throughout Virginia, Washington, D.C. and Maryland. Crestar Mortgage Corp. also serves customers throughout the southeastern U.S.

How can a company learn about its customers? Where do they live? What products do they own and use?

These are the types of questions that the market research staff at Crestar Bank have been answering for the last few years. Rick Kolster, consumer market research manager at the Richmond, Va.?based bank, explained that in order to answer such questions, certain information tools are needed.

To develop these tools, Crestar's consumer research group first looked at the information within the bank's own customer information system. Selected account information for the entire customer base was gathered from many separate data processing systems and integrated in a uniform file. By making a copy of the entire file and adding to the information other pieces of data from sources outside the bank, a database was created that effectively is being used as a market segmentation system.

Data elements

Kolster explained that the building of the system started with the most basic pieces of information such as customer name, residential address and Crestar product ownership and usage. Since certain financial relationships are better understood at the household level rather than the individual customer level, specific customer information was processed in such a way as to identify those customers who comprise a single household.

The process, known as householding, is a computer-assisted match routine that focuses on the customer last name, residential street address - both number and name, and the zip code. Each identified household is assigned a unique number, just as individual customers are assigned a unique number. Having the households identified allows the research staff to associate Crestar product ownership at both the individual customer and household level so that a complete relationship with the bank can be identified.

Knowing which bank customers make up households is valuable by itself but Crestar has added information to its segmentation system that permits the research staff to know more about the nature of the households that make up its customer base. The additional insight is provided by the use of a geography and Census demographic information system that provides data elements for the segmentation system.

Cluster codes

The information that has been added is known in the market research industry as cluster codes. A number of market research vendors offer such systems, packaged in a number of different forms ranging from "stand alone" desktop computer work stations to data that are run on mainframe computers. Regardless of which form is used, the basic element of the respective cluster code systems is the format for the codes themselves.

Crestar uses the codes in both a desktop work station and on the bank's mainframe computer. The particular system Crestar uses considers 117 demographic, socioeconomic and housing characteristics from the 260,000 U.S. Census block groups to identify 48 homogeneous cluster segments. Each Census block group is assigned to one of the 48 segments, making it possible to draw conclusions about the nature of households based on the residential address and the identified Census block group.

Each of the cluster codes are described by a written description that characterizes the dominant age of the household head, likely household income level, presence or absence of children, age and estimated value of housing.

Segmentation system

Thus, Crestar's market segmentation system is a database that combines bank customer information, identifies households, and identifies Crestar product ownership and usage. In addition, standardized geography codes, based on the U.S. Census, are used to permit inter- and intra-area comparisons of household types using characterizations found in the cluster code system.

The information is used in a number of different ways. The most basic use is a profiling of the customer base in various ways, such as geographic concentrations, product ownership concentrations and measuring the strength of the market as shown by types and numbers of households. Having this type of information provides answers to some of marketing's most basic questions. However, simply having a "picture" of the customer base is not enough. The information has to be used in actual marketing activities to make the system truly valuable.

Information at work

Once the research staff knows where the customers are located, judgment can be made as to how effectively the bank serves its customers with the branch system, promotional appeals and media messages. Due to the use of Census- based geography, measurement can be made of areas within and outside the trading area the bank considers as its markets. For key segments to which the bank specifically wants to give attention, the cluster codes provide a way to measure the size and know key attributes of the market.

The segmentation system allows the research staff to know how well Crestar products have been sold in the past and the configuration of product ownership within types of households (cluster codes). Having the Crestar households identified with cluster codes permits examination within the customer base and also provides insight as to financial product ownership for those households which do not use Crestar products but are assigned the same cluster codes. This information becomes the basis for making promotions to customers with whom the bank would like to develop a relationship.

Kolster speaks highly of the use of cluster codes in terms of its ease of use and timeliness. Considering their use and other Census information in a desktop work station, Kolster says, "It gives us the convenience of having a tremendous amount of information at our fingertips. The information is easily accessible, timely and gives current estimates for forecasting and making projections. Since this information can be displayed on a map, the system makes communicating information easier, too."

Better decisions

Kolster says the reliable information which the segmentation system generates has enabled Crestar to make "bigger and better decisions with greater certainty. Our objective is to know the customer base and identify those customers who are strategically important to us. We also want to know why they chose the product mix that they have and ultimately, how we can make our customer base grow. Simply, the segmentation system is a way to help decide how to make efficient allocation of our resources."