Ten Steps to Data Quality

by Danette McGilvray

May 2013

“I’m a busy manager. I have a business to run. Why should I care about data quality?”
”Data Quality? I have an important technology project in progress. Why should I care about that?”

You have data quality whenever someone in your organization pulls up a customer name and address to take an order, process a shipment, send marketing materials, or create an invoice and you can depend on that information being correct. You can count on it to efficiently complete the business transaction. Though it may seem obvious, let me point out what that means to your customer - the one who provides the revenue that keeps your company in business. The customer is happy that the order was taken quickly because the correct information was available and the representative confirmed that the desired product was in stock. The shipment was received by your customer with no problem because the shipping address was correct. The invoice in the package was correct and matched what the customer expected when the order was placed.
Data quality also means you can trust your business intelligence and analytics. Whenever you bring together information about your top customers across the enterprise, you can confidently adjust your strategy based on those reports. If your staff is spending time arguing about whose report is correct (when the numbers should agree, but do not) - you have a data quality problem. How much time is wasted trying to reconcile the reports instead of taking action based on what is learned from the reports?
Once again, why do these things matter? The more effectively we take care of our customers the less it costs the company. This also increases the likelihood our customers will do business with us again, which then increases revenue.
If you are implementing a new system, such as a Customer Relationship Management (CRM) or ERP (Enterprise Resource Planning), or replacing a legacy application with the latest technology, you are most likely integrating and moving data from various source systems. If so, I guarantee you will have a data quality problem.
I can say this with confidence because every project I have personally been involved with or heard about that integrates and moves data has had data quality problems that had to be addressed for the application to work correctly when put into production. Data that may have fulfilled the needs of one particular functional area are now combined with data from other functional areas – often with very poor results. Data quality problems often catch technology projects by surprise. Whether you are just starting your project or are already in production, it is not unusual to find that information and data quality issues prevent the company from realizing the full benefit of their investment in the new systems. This is the bad news. But the good news is there is something you can do about it to avoid the problems that poor data quality brings.
Just as there are people in your company with specialized knowledge in technology (data architects, application developers and programmers, etc.) and business areas such as accounting (Chief Financial Officer, controllers, accountants, and bookkeepers) there is specialized knowledge available related to data and information quality. One source for help to address data quality challenges in projects is a book called Executing Data Quality Projects: Ten Steps to Quality Data and Trusted Information™ (Morgan Kaufmann) by Danette McGilvray.

The Ten Steps™ methodology is made up of a framework, concepts, and processes for creating and improving information and data quality within any organization. The Ten Steps process contains concrete instructions for executing information and data quality improvement projects. To implement The Ten Steps process effectively, it is necessary to understand some concepts related to data quality; for this reason a Framework for Information Quality and several key concepts are presented in the early chapters. Readers are given enough background on the underlying concepts to understand the components necessary for information quality and to provide the foundation for the step-by-step instructions. The instructions provide enough structure for readers to understand what needs to be done and why. The beauty of the approach is that it provides just enough structure to know how to proceed, but flexibility so those using it can also incorporate their own knowledge, tools, and techniques. It was written to fill the gap between, yet be a complement to, existing books that provide higher level concepts or processes and other books that dive deep into specific data-related subjects.

• Organizations can choose the applicable steps, activities, and techniques for their situation and use the methodology:
• For information quality-focused projects, such as a baseline data quality assessment of a particular database or a business impact assessment to help determine appropriate investments in data quality activities.
• To integrate specific data quality activities into other projects and methodologies, such as building a data warehouse, implementing an ERP (Enterprise Resource Application) or migrating data for any application development project.
•  In the course of daily work or operations, whenever you are responsibility for managing data quality or the work you do impacts data quality.
• As a foundation for creating your own improvement methodology, or to integrate data quality activities into your organization’s standard project life cycle or software/solution development life cycle (SDLC).

Use of this approach can save an organization time and money because the foundational work for the process has already been done. Therefore, your time is spent determining how to make it work for your specific situation, taking action, and getting results.

If your challenge is to show why data quality is important and gain support for the resources needed, the methodology includes business impact techniques. These are qualitative and quantitative measures for determining the effects of data quality on your organization. These techniques range from less complex (such as collecting examples or stories about the impact of poor data quality) to more time consuming techniques (such as quantifying the costs and revenue impact of poor quality data). Having real results of business impact combined with good communication skills will go a long way towards getting support for your data quality efforts.

Though I have used customer information as examples in this article, what I have shared works with all types of data - financial, employee, product, manufacturing, medical, research, and so forth. Whether your organization is a for-profit business, a government agency, a charity, an educational institution, or related to healthcare, all of these ideas apply – because every organization depends on information to support its goals and to deliver on its commitments. If your company is concerned about data quality, now is the time to get started!