The current generation of customer data integration (CDI) and master data management (MDM) software provides powerful tools for managing data quality. There is, however, still a significant human element to the overall quality of your data. For example, a wide range of people may be entering data about entities. Salespeople may enter identity and transaction data about leads and customers, customer service people may enter information about new individuals within customer companies as well as transaction data, and order entry people may enter customer identity information. Similarly, people throughout the enterprise may modify, update and enhance data for sales, marketing, customer service and billing needs. All this disparate data - created by many people over different time periods - can be aggregated and partially cleaned programmatically, but maintaining high quality data over time will require implementing the five keys to data quality.
Number 1: Establish data standards and definitions. Standalone data systems generally have implicit, and sometimes explicit, data standards and definitions. The users of the system have agreed on standards or definitions in order to make the system work. These standards and definitions are almost certainly not consistent across all the data systems within your enterprise. They need to be made explicit, and people across your organization need to agree to them. The last sentence sounds significantly easier than the actual practical experience. People resist change, and that includes changing practices about standards and definitions. One compelling reason for recruiting a strong executive sponsor as well as a cross-functional steering committee of senior managers for your project is to provide the leverage needed to make these changes.
As you work to establish standards and definitions, take the time to determine which definitions and standards were created or evolved for internal reasons (convenience and convention) and which were developed because they efficiently reflect external reality. The former are much easier to change than the latter.
Number 2: Create repeatable processes for data collection, enhancement and maintenance. If you regularly read my column, you already know that the two primary enemies of entity databases are the complexity of companies and rate of change of identity information. The implication of these two external realities is that your data quality efforts are never finished. As you develop your quality enhancement processes and metrics, you will need to ensure that they can be performed in the same way over and over again.
Number 3: Monitor performance at multiple points in the transformation process. Data quality can be measured relative to business end-user needs, relative to alternative data sources and relative to prior periods. If you are managing quality strategically, you will be focused primarily on end-user needs. But waiting until the data gets to the end user is often too late to correct data problems. Successful data quality programs generally measure quality - usually relative to prior periods - at each step in the transformation process. If the data quality enhancements have been selected to drive improved end-user satisfaction, these intermediate metrics should be good predictors of your ultimate success.
Number 4: Implement continuous process improvements. We have all heard people say that they have plans to improve data quality dramatically, but they just cannot find the funding. I have never seen a case where implementing a continuous improvement process - a recurring process looking for opportunities to eliminate, consolidate, standardize and then automate - would not free up enough expenses to pay for the desired data quality enhancements. Eliminating, consolidating and standardizing across an organization is never easy. The role of a strong executive sponsor, supplemented with a steering committee of senior executives, may be needed to drive the process. Just because you have always done things one way does not mean they need to continue to be done that way, especially if sticking with current practice takes so much of the expense budget that you cannot rationalize, clean and maintain your customer and prospect data.
Number 5: Hold people accountable. It is probably safe to assume that everyone involved in creating and maintaining customer and prospect data intends to do a good job. But they probably have lots of other tasks that need to be done, too. If you count on good intentions to lead to better data quality, you will probably never reach your goals. The people responsible for creating, rationalizing, cleansing and maintaining data need clear metrics for their activities so they can understand what they are expected to do. Probably more importantly, they need personal goals against those metrics, and they need to be rewarded for achieving and exceeding their goals. Once again, a strong executive sponsor may be necessary to ensure that compensation plans across the organization are modified to include data quality goals.In my experience, the behaviors of people involved in data management and quality enhancement programs are at least as important, and may be more important, to data quality than the software choices that you make. Use these five keys to help you organize the people processes.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access