Almost by definition, a customer-centric strategy demands identification of each unique customer within the customer community. Creating a representative model of the customer is a necessary prelude to developing customer profile models and analyzing any characteristics and behaviors for classification. That model must, at the very least, incorporate these aspects:

• Description of the entity – namely an explanation of what the business roles are for the represented entity, how those roles are manifested within business processes and how those roles are distinguished within the community. For example, there may be a concept of a “head of household customer” who is responsible for payment for a provided service, while there may be other “household member” customers who do not have financial accountability but are eligible for customer service.
• Identifying characteristics, which comprise the collected set of data attributes that can be used to uniquely distinguish one customer from another. An example of a set of identifying characteristics might be first name, last name, date of birth, city of birth and state of birth. In addition, there should be a way of differentiating identifying characteristics from the behavioral or descriptive characteristics that will contribute to the customer analysis. These are typically the data elements whose data values are inherent to the individual.
• Unique identifier, which is a key value that is associated with one, and only one customer.
• Methods for recognizing entity information, which are critical in dissecting individual data embedded within semi-structured data or for recognizing entity information embedded within unstructured data. As an example, account-oriented data models often subsumed multiple identities associated with a single bank account’s name, such as “Mr. and Mrs. John Franklin UGMA John Franklin Jr.” In this case, there are three unique individuals referenced in the account name: Mr. John Franklin, Mrs. John Franklin and John Franklin Jr.
• Methods for resolving unique identity, which become necessary as the customer community expands. Initial attempts at determining identifying attributes may work well until the pool of individuals expands to include people who share their set of identifying attribute values. At that point a process must be used to differentiate between two real-world entities that share the same set of identifying attributes. This may require the augmentation of the set of identifying characteristics, which may require a recalculation/repopulation of the customer database.
• A customer entity data model. This model must not only be able to capture the identifying characteristics, it must enable linkage with additional entity data (such as locations or contact mechanisms) as well as subsume relationship and hierarchy connections that are used to help evaluate any individual customer’s sphere of influence.