This often happens because the sales organization defines "customer" one way, the finance team defines it another and the customer service group begs to differ with both of them. The end result of this difference (in opinion and in systems) is fragmented customer data, leaving a critical corporate asset being defined and managed differently in different parts of the business, with little or no systematic integration between the different silos.
This can be crippling to the process of becoming more customer-centric, increasing profitability and to making good day-to-day business decisions.
In today's more complex business enterprises, companies need to put an organization and process in place, supported by the appropriate technology and trusted external content, to manage master data across the enterprise in order to increase revenue, reduce costs and improve compliance.
This holistic combination of organization, processes, technology and information is termed master data management (MDM), and the techniques and approaches, as applied to the "customer" domain within the enterprise, are known as customer data integration (CDI).
Defining the Space
MDM is a set of disciplines and processes for ensuring the accuracy, completeness, timeliness and consistency of the most important types (or domains) of reference data in the enterprise - across different applications, systems and databases, and across multiple business processes, functional areas, organizations, geographies and channels.
The most common MDM domains are customer data, known as CDI in the MDM subset, and product data, known as product information management (PIM) in the MDM subset.
Critical MDM capabilities include data quality, identity management, data enrichment, grouping, synchronization and process management.
Data quality capabilities are critical to ensure that only accurate, completely, timely and consistent data is being synchronized across the entire enterprise.
Identity management serves three important and related functions:
- Recognize an organization or person (and whether they're a prospect or an existing customer) regardless of which channel (call center, Web store, etc.) they use to interact with you,
- Validate identity so you're confident people are who they say they are, and
- Prevent duplicates, so you avoid adding a new record unless it's "truly" new.
Data enrichment brings in external information on an organization or person, telling you valuable things you didn't already know.
Grouping links organizations and persons in useful ways, typically in corporate hierarchies when dealing with businesses and households when dealing with consumers.
Synchronization and process management (usually through middleware and business process management software) allow you to not only move information from Point A to Point B in the enterprise, but also to do more sophisticated things than simply move information. You can have processes that are long running, tightly monitored and controlled, and span multiple applications and even enterprises.
Without a systematic way to manage critical master data, collaboration across the enterprise, between the diverse IT systems and the various business functions, can be difficult and costly. But, as MDM becomes more widely used, organizations are starting to figure out how to build robust MDM solutions, using a combination of off-the-shelf hubs, middleware and process management tools, plus data quality software, Web services, service-oriented architecture and custom components.
It is evolving this architecture that will prove critical in the future as organizations need better information, increased agility, efficient processes and less costly compliance to compete in today's increasingly "flat" and competitive business environment.
Dealing with Dirty Data
MDM by its nature is tough, partly because a company often expects it to solve some of the "holy grail" problems and because of the number of moving parts and the state of the market. The hard things are easy and the easy things are hard - the organizational and process questions turn out to be the toughest, and the technology and information challenges, while not trivial, can be managed.
One of the biggest of these challenges is data decay. D&B manages a global commercial database with more than 110 million business records and finds that, i n one year:
- 21 percent of CEOs will change
- 20 percent of all addresses change
- 18 percent of telephone numbers will change
- 17 percent of business names will change.
It almost doesn't matter which data element you're talking about - approximately 20 percent of the values in that field will change over a 12-month period. Most people assume that customer data is fairly static, and it just has to be cleaned centrally one time before being distributed throughout the enterprise. Unfortunately, that's a big oversimplification.
The truth is that collecting information is not the hard part. It's maintaining and evolving the data over time that is difficult. As business needs change, the "single view of the customer" has to evolve and change with it. Otherwise, the promise of MDM and CDI will be unfulfilled, and today's hubs will be just another brittle silo of customer data within the enterprise.
Businesses today also want more and different ways to look at their customer data - "functional area" views. For example, sales may want a way to view customers by assigned salesperson and roll it up in a "sales organization" hierarchy; marketing wants a way to view customers by rolling them up by customers' legal hierarchy and customer service is looking to quickly find the existing customer record no matter what channel that customer comes in through (phone, Web, in-person visit, etc.) and no matter what information the customer provides (account number, phone number, email address, etc.).
