THE DATA ADMINISTRATION NEWSLETTER – TDAN.com
ROBERT S. SEINER – PUBLISHER

Is Information Quality Important for Business Intelligence?

It is no secret that Quality Information is a critical success factor in Data Warehousing. After all, what is the product of the Data Warehouse? A Data Warehouse is dangerous to the health of the enterprise if it delivers inaccurate, missing or poorly-defined information and knowledge workers make the wrong decisions. Take two examples. A global financial institution invested in a company thought to be credit-worthy. It ended up losing more than one-half billion dollars ($500 million+) because of differences in the definitions of Risk codes across its business units led to the wrong decision. A major bank lost over $200 million because changes in the credit rating information received from its Credit Bureau caused it to send out pre-approved credit cards to thousands of individuals who were not credit-worthy.

The good news: both organizations are on sound journeys to implement a mature pro-active IQ management environment.

The High IQ Journey

There is nothing magic about getting to a high Information Quality environment. But there are no silver bullets. There is no software product you can buy and drop in and, "Presto!" Instant IQ. Success requires a disciplined approach that implements sound quality management principles, processes and practices to make IQ a part of the culture.

Here I address the 5 Stages of Information Quality Management Maturity (IQMM™), and how to embark on the journey to increase your IQ[i] and enable an Intelligent Learning Organization. The IQMM is based on Philip Crosby's Quality Management Maturity Grid,[ii] which was also used by the Software Engineering Institute as one of the inputs to create the Capability Maturity Model (CMM) for software quality.

There is much being written about Information Quality and Information Management Maturity today, but unfortunately much of it is not well founded, nor is it correlated with demonstrated maturity.

Organizations implement IQ in different ways. Some create effective IQ environments while others sub-optimize their IQ effectiveness. Those differences will separate the Data Warehouse throw-aways from the competitive, High-IQ enterprises. The Five Stages of IQ Management Maturity

Here I address the 5 Stages of Information Quality Management Maturity (IQMM™), illustrated in Figure 1, that have been adapted from Crosby to reflect the same maturity stages applied to Information Quality Maturity[iii] and how you can embark on a sound journey to maximize the effectiveness of your Information Quality Management function. They include the following and the general behaviors exhibited in them:

Stage 1: Uncertainty. "We don't know why we have problems with information quality."[iv]

This stage represents the clueless organization. The Data Warehouse will be populated with data without any or with minimal data correction processes.

Figure 1: The Information Quality Management Maturity Grid[v]

Stage 2: Awakening. "Is it absolutely necessary to always have problems with information quality?"

The Awakened organization has become aware of the problems caused by poor quality. They generally attack IQ problems by correcting data in the data warehouse or in the staging area before loading. Here, no efforts are made to solve the real problem of poor quality data at the source. If any are made it is only to correct data there, then propagate the data to the Data Warehouse.

But, from a quality management perspective, if an organization only conducts data cleansing, it merely automates "information scrap and rework."[vi] To sustain a quality system, you must move from data cleansing only (corrective maintenance) to process improvement (preventive maintenance). Process improvement leads you into Stage 3.

Stage 3: Enlightenment. "Through management commitment and information quality improvement we are identifying and resolving our problems."

The Enlightened organization recognizes the cause of poor quality information is broken processes and systemic performance measures that reward speed and departmental objective accomplishment even if it sub-optimizes down stream processes. Business executives become involved and drive the IQ initiatives.

In this stage, the organization is implementing a formal quality management system, such as the 14 Points of IQ,[vii] Deming's 14 Points of Management Transformation,[viii] Kaizen,[ix] the Baldrige Criteria for Performance Excellence (that now includes Information Quality),[x] or Six Sigma and applying them to information.

Enlightened organizations conduct data cleansing as a one time event for a given data source and will improve the processes at the source where the defects are caused, using a sound process improvement technique like the Shewhart Cycle, or PDCA (Plan-Do-Check-Act):

• Plan an improvement to prevent recurrence of the defects by first analyzing root cause (not just precipitating cause) and defining improvements that prevent recurrence.
• Do implement the improvement in a controlled environment.
• Check to confirm it has solved the problem.
• Act to put the process in control and roll it out to all of the enterprise.

For a description of how to conduct a PDCA for information process improvement see Improving Data Warehouse and Business Information Quality, Chapter 9 "Improving Information Process Quality: Data Defect Prevention."

In Stage 3, you have formal IQ training for managers and for information producers so they know who their information customers are and how to improve their processes.

The results of Stage 3 include huge reductions in the costs of process failure and information scrap and rework caused by defective information. These benefits become permanent because processes are now in control.

Stage 4: Wisdom. "Information quality problem prevention is a routine part of our operation."

Wise organizations are improving their quality management systems, and improving processes is a habit of the majority of employees. Every manager is accountable for the information produced or updated by the processes he or she oversees. All employees are routinely improving processes.

Stage 5: Certainty. "We know why we do not have problems with information quality."

The only nonquality information is that produced by processes outside the control of the enterprise, or that are temporarily caused by "information quality decay"[xi] caused when characteristics about real world objects change more frequently than is feasibly possible for the enterprise to become aware of immediately.

In Stage 5, called "Optimizing" in the CMM, the habit of process improvement is so engrained, that initial designs of process, applications and databases are designed with strong defect prevention mechanisms, and defects rarely occur.

Conduct an IQMM Assessment

Review the characteristics of IQ Maturity in Figure 1. Get together with knowledgeable people from across the organization and conduct an IQMM self-assessment. Be honest. Fudging only gives you a false sense of security.

Ask yourself which stage of maturity is required to prevent your organization from being victim to the kinds of problems illustrated at the beginning of this article. Then:

1. Take action to learn quality management principles and apply them to your information processes with disciple.
2. Identify a critical set of information where the risk or costs of nonquality are great both at its source and in your data warehouse.
3. Perform a Proof-of-Concept IQ project:
   • Measure the cost of nonquality before you correct the data and improve the process to document the high costs of nonquality information.
   • Conduct a one-time corrective maintenance (cleansing) of the data at the source and,
   • Conduct a PDCA to improve the processes to eliminate the cause(s) of nonquality and put it in control.
   • Measure the cost of nonquality after, so you document the return on investment.

And as you mature, you will experience the exhilaration of increased value of your information, increased effectiveness of your business and systems process, increased knowledge worker satisfaction. And, most importantly, increased customer satisfaction, along with increased profit and decreased costs of operations.

Conclusion

Organizations with a mature IQ environment (Stage 3 and beyond) are the ones who will survive and thrive in the realized Information Age. These organizations will redefine the economics in their industry in the same way the Japanese redefined the economics of automobile and consumer electronics manufacturing.

To a High IQ!!!

[i] The double entendre here is that High IQ (Information Quality) leads to a High Intelligence Quotient for the enterprise, or the Intelligent Learning Organization.

[ii] Philip Crosby, Quality Is Free, NY: Penguin Books, 1979, see pages 21-136.

[iii] For a detailed description of the IQMM, see Larry P. English, Improving Data Warehouse and Business Information Quality: Methods for Reducing Costs and Increasing Profits (IDW&BIQ), NY: Wiley & Sons, pp. 427-437.

[iv] Ibid., p. 428.

[v] Ibid., p. 428.

[vi] Ibid., pp. 210-212.

[vii] See English, IDW&BIQ, chapter 11, "The 14 Points of Information Quality," pp. 337-399.

[viii] W. Edwards Deming, Out of the Crisis, Cambridge: MIT Center for Advanced Engineering Study, 1986, pp. 18-96.

[ix] Masaaki Imai, Gemba Kaizen, NY: McGraw-Hill, 1997.

[x] Baldrige National Quality Program 2004 Criteria for Performance Excellence (Business version), accessed April 7, 2004 at http://www.quality.nist.gov/PDF_files/2004_Business_Criteria.pdf

[xi] English, Op. cit., p. 205.