设为首页 | 添加收藏
首 页 走进石竹 石竹动态 产品中心 解决方案 服务体系 知识中心 招贤纳士
 
K 知识中心  
数据治理
元数据
数据质量
技术讲坛
下载
 
C 联系方式  
北京:(8610)58810139
广州:(8620)61246699
上海:(8621)62351222
销售联系:
info@carnation.com.cn
技术支持:
support@carnation.com.cn
 
 
  当前位置:首页 >> 知识中心 >> Total Information Quality Management ¬– A Complete Methodology for IQ Management
Total Information Quality Management ¬– A Complete Methodology for IQ Management
发布日期: 2017-04    阅读次数: 7488

·                 Larry English

·                 DM Review Magazine, September 2003

A colleague of mine recently told me that he is working for a consulting firm on an information quality (IQ) project with a client where they were just "playing with data quality," not really doing it. As he shared some details, we both agreed it was a sad state of affairs. This article explains what is required to really do information quality and describes a methodology for how to do it.

It would be really nice if we could solve IQ problems by implementing a one-time project. However, IQ problems cannot be solved that way because IQ problems are caused by broken, out-of-control processes, which are influenced by industrial-age, ineffective management systems that reward the wrong things.

It would also be nice if we could solve IQ problems by simply dropping IQ software systems into place to completely clean the defective data. Alas, IQ problems cannot be solved that way either. Software cannot completely remediate the limitations and weaknesses of broken processes outside the software system, nor can it completely eliminate the factors introduced by the human element in information production. IQ problems cannot be solved permanently unless you address the organization's management systems that set performance measures and influence employee behavior.

There is no panacea or silver bullet for Total Information Quality Management (TIQM), just as there was no panacea or silver bullet for manufacturing quality management. TIQM requires a sound, defined set of processes ­ implemented and executed with discipline.

Importance of Methodology

Five Reasons to Implement TIQM

Every process that produces quality results consistently requires the process to be defined (so that it is repeatable), controlled (so that it is consistent) and improved (so that it eliminates the causes of defects or error). This is true whether the process is manufacturing a car, filing a tax return, taking a catalog phone order or processing an insurance claim. This is also true whether the process is developing an application system, developing an enterprise data model, conducting an IQ assessment or facilitating an information process improvement.

1. TIQM was developed using proven quality management principles, methods and techniques successfully applied to manufacturing and other industries.

When I began my work in IQ management, Tom Redman had not yet written his first book (published in 1992). There was no formal program of IQ at any university. I became interested in information quality by discovering W. Edwards Deming ­ first through Mary Walton's book, The Deming Management Method. From that, I moved to Deming's Out of the Crisis in which he outlines his famous 14 Points of Quality.

Deming's 14 Points transformed my professional focus and consulting practice. I recognized immediately that quality management principles are not ancillary to information; they are directly applicable to data as the product of business processes in the same way an automobile is the product of manufacturing processes. Then I read Masaaki Imai's Kaizen: The Key to Japan's Competitive Success. It confirmed what I learned from Deming: Quality management is a continuous improvement of processes to eliminate the causes of defects. Another major influence was Philip Crosby's Quality Is Free: The Art of Making Quality Certain. It confirmed again that the business case for information quality is the same as the business case for quality management in general ­ measure and eliminate the costs of scrap and rework. I also learned from Crosby the five distinct stages of organizational maturity in the journey and transformation to an effective culture of quality management and business performance excellence.

In addition to these quality gurus, I studied Joseph Juran (Juran on Planning for Quality), Kaoru Ishikawa (Guide to Quality Control), Yoji Akao (Quality Function Deployment: Integrating Customer Requirements into Product Design) and the Baldrige National Quality Program's Criteria for Performance Excellence. Now I am studying Six Sigma and have mapped TIQM's process steps into Six Sigma's define-measure- analyze-improve-control (DMAIC) methodology.

TIQM was not developed by looking at other data quality or information quality methods and techniques, but by studying quality management principles, techniques and processes from the leaders of the quality management revolution, understanding them and then applying them to information.

If you doubt the pertinence of manufacturing quality management principles to information, please note that before Deming went to Japan and taught the Japanese his quality management principles, he applied them at the U.S. Census Bureau in the 1940 national census process ­ a pure information quality application!

2. TIQM is IQ software tool open.

TIQM is not biased to any specific software tools. Rather, it provides a framework for exploiting IQ software as well as other quality tools and techniques. It classifies the capabilities of IQ software and where they fit in the TIQM process methodology. Tool categories include: assessment; rule discovery and analysis; reengineering, cleansing and transformation; enhancement; defect prevention; and quality control and management.

You may ask, "Why should our company implement TIQM? After all, we have implemented IQ software/our own methodology."

First, if you already have an IQ methodology, compare it to TIQM and augment it if necessary. Remember that every process is a process that can be improved (Deming's Point 14) ­ even IQ processes.

Many of you may have excellent IQ software in which the supplier has provided a methodology. Do not throw that methodology away, but examine it against TIQM to see if there are gaps. Before I began IQ consulting, I noted that when third-party software suppliers developed methodologies to augment their products, those methodologies were biased based on the features of each supplier's products. Important methodological steps were missing if the product did not provide a function to support that step. The same omissions can happen in IQ software providers' methodologies. For example: IQ assessment software suppliers may provide a methodology for IQ assessment that includes steps to define, measure and report completeness and validity assessments. Does the methodology have a step to measure accuracy by physically comparing data to the real-world object (or a recording or observation of an event)? Accuracy assessment, one of the most important IQ characteristics for knowledge-workers and end customers, cannot be conducted electronically. Comparing data to a reference database, such as a postal database, does not represent a measure of accuracy; it represents a measure of consistency. In order to interpret the result of this comparison, you must know the accuracy of the reference data. For example, an address may be determined to be a valid address, but the person believed to live there has long since moved.

To measure accuracy requires one to actually compare the data to the real-world object or event the data represents to confirm the correctness of the data values to the real-world object characteristic represented. TIQM describes how to conduct a physical accuracy assessment step depending on the different object types, such as person or organization, physical object, location, intangible object or event.

If your methodology only measures validity and not accuracy, you may actually have a false sense of the real level of quality of your data. Your data may meet every validity measure but be inaccurate to a costly degree. For example, a U.S. federal agency acted on "reasonable" data in providing funding. An audit showed it misallocated 11 percent of its funds in one program in one year, including making $1.7 billion ­ not million ­ in excess payments and underpaying an additional $600 million entitled by worthy recipients!

Data cleansing software suppliers may provide a methodology for cleansing that includes steps to correct the data only as it is moved from the source to a target database. Does it include steps to update the data at the source databases if the data is still being used there? Failure to correct data at the source creates a new IQ problem of inconsistency of data in the target and data in the source. It also causes processes using the data at the source to continue to fail because of the uncorrected data. Furthermore, the methodology may omit altogether the process of Plan-Do-Check-Act (PDCA) to analyze root causes of the defective data that requires cleansing and the process improvements required to prevent recurrence of the defective data.

If your methodology only cleanses data but does not improve the processes, it actually institutionalizes the information scrap and rework of "continuous data cleansing" rather than "continuous process improvement." Take my earlier example of $1.7 billion dollars in overpayments. Data cleansing by itself will not prevent the continuation of misallocating funds. TIQM defines a process for process improvement.

Even IQ software suppliers that provide defect-prevention capability may have a methodology with steps only for implementing the defect- prevention capabilities of their tools but may completely miss inclusion of steps to analyze process improvements for the overall process such as form and screen design, manual procedure and error- proofing techniques that minimize human error.

Does this mean that you should not use an IQ software supplier's methodology? No. Software providers' methodologies tend to be optimized for the capabilities of the tool. At the same time, those methodologies may be limited by the tool capabilities. Maximum benefit is achieved when you implement tool-oriented methodologies in the context of an umbrella IQ management system, such as TIQM. This assures you have a complete methodology that addresses the full scope of IQ management and optimizes the capabilities of the IQ software.

3. TIQM is a complete methodology for information quality management.

TIQM includes a process to measure the costs of nonquality information and then the value of improvement. This is the business case and may be the only measurement that management cares about. What is the business motivation from the fact that the product database contains 7.3 percent errors? The real business case is that those 7.3 percent errors cost the company $9.3 million dollars in lost sales, cause $3.4 million in compensation to alienated customers and cause lost customer lifetime value of $15.2 million.

TIQM defines a process for process improvement that applies PDCA to information processes. In fact, TIQM's Six Processes and their steps have been mapped to Six Sigma's DMAIC methodology.

TIQM defines the principles for the culture change required to accomplish the transformation of organizations from vertical function management to horizontal value-chain management. Furthermore, it provides a road map for you to tailor to your unique situation to begin and sustain your journey toward information quality and business performance excellence.

4. TIQM defines a process for establishing the business case for IQ.

TIQM provides a process to measure the most significant aspect of nonquality ­ its cost to the business in the form of process failure, information scrap and rework, lost customer lifetime value and missed opportunity. When you need to make a business case for management to invest in IQ, conduct a P3 initiative (see TIQM process P3).

5. TIQM is supported by documentation and training.

TIQM is described in my book, Improving Data Warehouse and Business Information Quality, published by John Wiley & Sons. Part II describes each of the five discrete TIQM processes, listing the process steps and purpose, inputs, outputs and activities of each process step. TIQM Process 6, "Establish the Information Quality Environment," described in Part III, includes a detailed description of the 14 Points of Information Quality, adapted from Deming's 14 Points of Quality, along with a description of the steps to implement an effective information quality environment.

Detailed training is available for all aspects of TIQM, including detailed training for information quality practitioners, training for business management ("IQ Management: What Managers Must Know and Do"), and training for information producers and knowledge- workers to help them understand how to improve their processes and error-proof their work.

Some organizations have licensed TIQM education and have certified instructors to teach it within their organizations.

The Six Processes of TIQM

TIQM consists of five discrete processes of measurement and improvement and an umbrella process of implementing a cultural transformation to grow an environment of valuing information customers, a mind- set of process excellence and a habit of continuous process improvement. This figure illustrates the Six TIQM Processes and their relationships. Note: The methodology is not sequential; you may start at any process. The first three processes are assessment processes. P4 is a process of defective data improvement (data correction) and data movement control. P5 is a process of process improvement to prevent recurrence of defects. P6 is a process of culture transformation.



Figure 1: The Six Processes of TIQM

P1. Assess Data Definition and Information Architecture Quality

This process defines how to measure the quality of data definition to meet the requirements of the knowledge-workers to know what they need to know and to understand the meaning of the information they require. This process also defines how to measure the quality (stability, flexibility and reusability) of data models and database/data warehouse designs.

P2. Assess Information Quality

This process defines how to measure the quality of information to meet the various quality characteristics, such as accuracy, completeness, non-duplication and consistency across multiple databases, as required by information customers. This process measures either the state of IQ within a database or data collection or the IQ produced by a current process.

P3. Measure Nonquality Information Costs and Risks

This process defines how to establish the business case for information quality management. By measuring the costs of process failure, information scrap and rework, alienation of customers and lost business, and missed opportunity, you can "speak with data" (as Kaoru Ishikawa says). That is, you can present this information in business terms to business management.

P4. Reengineer and Correct Data

This process defines how to conduct data correction projects, transform information and control data movement processes for data warehouses or data conversion projects. The process of data correction is not a standalone process, but one to be conducted along with a P5 initiative to improve the processes to prevent recurrence of the defects having to be corrected.

P5. Improve Information Process Quality

This process is required to call a methodology a "quality management" methodology. This process implements the Shewhart Cycle known as PDCA. It describes the tried-and-true process improvement technique defined by Walter Shewhart and used by W. Edwards Deming, Joseph Juran, Masaaki Imai (in the Kaizen quality system) and other proven quality management systems. It is through P5 that you deliver value to the enterprise by improving broken processes that cause defective data definition, data content or data presentation that, in turn, cause downstream process failure and high costs of information scrap and rework.

P6. Establish the Information Quality Environment

This umbrella step is not a discrete process with a defined beginning and end. It describes the 14 Points of Information Quality that must be inculcated into the culture of the enterprise to create and sustain an environment for business performance excellence and the habit of continuous improvement.

TIQM is a complete methodology for IQ management and improvement. Developed from proven quality principles and processes that have transformed the economics of the industrial age, TIQM can help transform the economics of the information age. Be part of the IQ revolution:

· Please study the TIQM poster. Read and understand its definition of information quality.

· Read about TIQM in the descriptive statements.

· Read and contemplate the 14 Points of IQ.

· Analyze the Six Processes of TIQM and evaluate them in the context of your environment.

· Apply the Six Processes of TIQM to increase your organization's IQ to eliminate the high costs of low quality information and increase your organization's process excellence.

Let me hear about your successes. Please contact me at LEnglish@infoimpact.com.

 

Larry P. English is president and principal of INFORMATION IMPACT International, Inc., Brentwood, Tennessee, and the author of the widely acclaimed book, Improving Data Warehouse and Business Information Quality: Methods for Reducing Costs and Increasing Profits. English is cofounder of the International Association for Information and Data Quality (www.iaidq.org). English is an internationally recognized speaker, teacher, consultant and author and may be reached at larry.english@infoimpact.com or through his Web site at www.infoimpact.com. For more on how to improve your IQ principles and techniques, and prevent your organization from wasting millions in information scrap and rework, join the IAIDQ (visit www.iaidq.org).

 

 
 
关闭窗口
 
石竹软件 版权所有 粤ICP备13017351号 | 公司月刊 | 招贤纳士 | 石竹质量 | 联系我们 | 邮箱登录 |
石竹北京 电话:(8610)58810139
石竹广州 电话:(8620)61246699
石竹上海 电话:(8621)62351222
销售联系:info@carnation.com.cn