Customer Data Integration: Reaching a Single Version of the Truth

DATA QUALITY AND MASTER DATA MANAGEMENT

Most companies that have taken on the data quality challenge have done so in a reactive and fragmented way. It’s easy to proclaim that “every customer must have an address.” But what if the customer record doesn’t include an address? Which address should be represented? We routinely meet finance executives who expect to see the customer’s billing address, call center reps who need the service address, and salespeople interested in the headquarters location. What constitutes a “bad” address can be relative. Few businesspeople have visibility into how their peers in different organizations need data delivered.

A casino we worked with decided to send out a campaign to frequent slot players. Reviewing the list of targeted customers, an observant marketing manager couldn’t help but notice that a large percentage of the customers on the list were 104 years old. Upon further inspection, the data team discovered that each time the date of birth field was marked as “Null,” the system inserted a “dummy” value to represent “Null.” The value was 1/1/0, or January 1, 1900. While there was one slot player who was indeed in her 80s, the other birthdates were invalid and had to be fixed. (And they’ve yet to meet the “high-value” 104-year-old craps player.)

Many people confuse data quality, data governance, and master data management (MDM). Given the definition of MDM in Chapter 2, we consider data quality to be a subset of MDM, which should support ...