This is the eleventh in a series of discussions of quality guru W. Edwards Deming's Fourteen Points of Quality and their ramifications on data quality.

Deming's seventh point of quality stresses management's accountability in quality improvement. Data quality will not be achieved solely by professionals measuring and reporting data quality or by "cleansing" data extracted from source databases before propagation to the warehouse. Data quality will be achieved when IS and business management lead the way to make process improvement happen and to remove the causes of data quality defects.

The job of management is leadership, not supervision, according to Deming.1 Management must not just supervise, but must continually work on sources of improvement of both product and service. Management's goal is to lead process improvement, not supervise workers to achieve quota. "Focus on outcome (management by numbers, management by objectives, work standards, meet specifications, zero defects, appraisal of performance) must be abolished, leadership put in its place."2 This statement is a bold contradiction to traditional management theory. Deming correctly identifies that focusing on quota and productivity numbers alone is, in fact, a major cause of quality problems. We see that emphasis on speed and productivity virtually guarantees data quality problems. A major mail-order company found that its goal to reduce time on the telephone caused a high incidence of duplicate customer records and unverified (and incorrect) addresses. A major stock trading firm found unacceptable errors in trade data. The culprit-- trader's bonuses were based on volume of trades, not quality of trade data.

Leaders are the primary enablers of improvement. Management must establish the environment for data quality improvement. Leadership means:

Knowing the work one supervises. A manager who does not know the work will be less able to coach workers to do the job correctly. Most measurement systems that managers put in place when they do not know the job encourage just the opposite, because the measurement systems reward the wrong things. For example, what is the relationship of function points and application quality? What is the relationship of "on time" and "within budget" to customer satisfaction?

Knowing common causes and special causes of quality problems.3 "Removal of common causes of trouble and of variation, of errors, of mistakes, of low production, of low sales, of most accidents is the responsibility of management."4 Leaders must be able to identify problem causes that are within the system or process (common cause). Common causes of data quality problems are those that can be corrected by process improvement. An example of common cause is the creation of a duplicate customer record because: 1) the procedure does not have a step to determine if the customer is on file, 2) the operator fails to look for an existing record, or 3) the automated algorithm is deficient and fails to find the customer's existing record. Special causes are those that are outside the system or process and are "special." An example of a special cause is the creation of a duplicate customer record because the computer or network system was down and the order was taken manually with contingency procedures. Special causes result in a process being out of statistical control. Removal of special causes enables a process to be stable, repeatable and predictable. Processes must first be put into control. Then they can be improved through the removal of the common causes of problems. Note that the contingency order procedure should have a step to discover and correct duplication, however.

Removing the barriers that prevent the staff from taking pride in their work, i.e., doing a quality job. Real leaders listen to their staff's suggestions for improvements. When workers understand who their customers are and what their customers need, they know how to improve the processes. If claims processors know that actuaries are their customers and they use medical diagnosis codes to analyze risk, they know how to improve the "create claim" process to meet customer quality expectations.

Establishing the right performance measures. Management often sets metrics that measure quality from an internal point of view, i.e., what management sees. Such measures as quotas, cost of production or speed of production without a measure of customer satisfaction are destructive. Even a measure of zero defects is wrong if it is not based on customer expectations.

Knowing how their group's processes fit into the goals of the enterprise. Management by objectives can cause different organizational groups to define conflicting and counterproductive objectives, even though they are supposedly decomposed from the same enterprise objectives. For example, a large publishing firm found that their order sales department and accounts receivable departments were constantly at odds with each other. The reason? Their objectives were in conflict. The order sales department had an objective "to increase sales" while accounts receivable sought "to reduce bad debt." Order sales took all orders to meet their objective even though some customers had risky credit. In turn, accounts receivable rejected some truly acceptable orders from marginal customers to assure they met their objectives.

Knowing how their group's processes fit with the upstream and downstream processes. This is what Peter Senge describes as "systems thinking"--seeing a part in the context of its interrelationships with the whole.5 Quality comes when all activities within the value chain work together to increase customer value. A leader "works in cooperation with preceding stages and with following stages toward optimization of the efforts of all stages."6

Knowing that half of any staff will perform "below the average" of the group. Trying to bring the below-average individuals up to standard is counterproductive and harmful. "Above average" performers will slow down and fail to achieve their full potential. Leaders provide resources to enable the staff to rise to their level of performance and not judge them based on the "norm." A person who does not perform quality work is "almost always in the wrong job or has very poor management."7

Building trust and providing help without judging. Leaders encourage everyone as a team to improve the processes.

Leveraging the skills of all. Leaders will encourage teamwork that allows experts to leverage their expertise by training and coaching others. The resulting team productivity is greater than the sum of individual productivity.

Leadership is required in both information systems and business processes.

IS Leadership

Information systems (IS) managers must provide leadership for information technology, application and database quality. IS leaders will see the interrelationship of technology, application and data components in an information-managed environment. They will further see that IS does not just supply services to departmental customers. Information systems must support optimization of the work across all stages of the larger business processes or value chains.

A more appropriate metaphor for the IS-to-business relationship is that of a partnership, not just a customer/supplier relationship. If IS sees its relationship as customer/supplier, it will have a tendency to deliver applications that meet the needs of one business area, while sub-optimizing the needs of other business areas.

IS leaders will:

  • Understand how their group's processes fit into the enterprise goals.
  • Understand the interrelationships of the three IS components of information technology, application and data.
  • Plan, model and build or deliver applications and databases across the business value chains. A business value chain is an end-to-end set of activities that begins with a request from a customer and ends in a benefit to a customer.
  • Work to improve the information planning and development processes. The key business information systems value chains are:
    • Information infrastructure development. This strategic process provides for the planning and development of the enterprise-wide information technology (network), business value chain and information (database) architectures.
    • Information systems component development. This operational process provides for the design, development of information technology, databases and applications as components of the architectures and their respective business and implementation value chains.
    • Employee development. Management of the employees not by numbers or standards, but by customer satisfaction and team-effectiveness criteria.

Business leaders will:

  • Understand how their group's processes fit into the goals of the enterprise.
  • Understand how their group's processes fit into the information value chain of which they are a part. This means knowing who are the downstream customers of the work and information products. It also means knowing the data quality requirements those downstream customers have of the information products produced by their group.
  • Understand that information is a strategic business resource and how quality information can be used to improve processes, decrease costs and increase new business opportunities.
  • Work to improve the business processes.
  • Manage the employees not by numbers or standards, but by customer satisfaction and team-effectiveness criteria.

How to facilitate leadership:

  • Be proactive.
  • Identify areas in need of improvement and in which you have influence; lead the team to make improvements.
  • Identify your customers and solicit their feedback on how to better meet their needs; then act on that feedback and get their new feedback.
  • Encourage and coach others to do the same.
  • Identify those who are visionaries and those who have data quality problems.
  • Listen to their problems and objectives.
  • Establish the business case for data quality from a perspective of how it enables them to solve their problems and meet their objectives.
  • Provide awareness education to management to sensitize them to the problems in operational efficiency and effectiveness caused by data quality problems.

Management by numbers simply maintains the status quo and focuses on arbitrary measures of performance and actually has the impact of decreasing productivity as a result of decreased data quality and the resulting costs of non-quality. Leadership provides vision, focuses on customer's expectations and implements continual process improvement to continually delight the customer.
1 Deming. Out of the Crisis, p. 54.
2 Deming. Out of the Crisis, p.54.
3 Aguayo, Rafael. Dr. Deming: The American Who Taught the Japanese About Quality, Simon & Schuster, New York, p. 76.
4 Deming. Out of the Crisis, p.321f.
5 Senge, Peter. Systems Thinking, p. 6-7.
6 Aguayo, Rafael. Dr. Deming: The American Who Taught the Japanese About Quality, Simon & Schuster, New York, p. 177.
7 Walton. The Deming Management Method, p.71.

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

Don't have an account? Register for Free Unlimited Access