Customer Data Integration - The Difficulty of Identifying Customers

The Difficulty of Identifying Customers

Given two sets of data that differ slightly, perhaps with a nickname ("Barb" instead of "Barbara") or with a cell-phone number, how can a computer distinguish between different aspects of the same customer and two distinct individuals?

In the larger sense, customer data can become very complex; for example, it may require anywhere from six to twelve or more fields to represent just a personal name. For example, some typical fields associated with a customer "name" may include:

• Name prefix (Mr., Mrs., Dr., Captain)
• Given name (a.k.a. "first name" or "Christian name" in some cultures)
• Family name
• Middle name(s)
• Name Suffix (Jr., Sr., II, III)
• Initials
• Nickname
• Maiden name
• Married name
• Professional title
• Academic title

Address entries have their own complexities (for example: primary address number, pre-directional (N, S, E, W,) street name, street suffix, post-directional, secondary identifier (building, suite, apt,) secondary number, city, state, ZIP, and ZIP+4(R))

Add in phone numbers, social security numbers, email addresses, tracking (customer number, account number), relationship, risk-level, purchase history, service history, demographic, socio-economic, lifestyle, consumer-behavior segmentation and privacy preferences, just to name a few.

This information varies constantly, whether due to changes, entry-errors, corrections, or fraud. And many medium-to-large organizations store this sort of data in several different places: different departments, different locations, in different formats, etc.