We are often asked what we consider the most important aspect of an association’s Business Intelligence initiative.   Based on our many years of experience, we find that Data Quality Management (DQM) is one of the most important and often overlooked components of Business Intelligence.

Why is it so important?

By definition, Business Intelligence is a process to improve performance by using data to make decisions, therefore, data is the foundation on which business intelligence is built.  Just as your house requires a strong foundation to ensure you have a sound home, Business Intelligence requires complete and clean data to ensure you are making sound business decisions.  If you make decisions based on incomplete, incorrect or duplicate data you run the risk of making bad decisions based on incorrect assumptions.  In fact the Data Warehousing Institute concluded that the cost of data quality problems exceeds $600 billion annually. What do you think the cost of data quality problems is for your association?

To some people, Data Quality Management might not sound as exciting as other parts of a business intelligence initiative.  However it must be tackled at the beginning of any business intelligence program to ensure you end up with valid information and a successful long-term analytics solution. Just like when building a home, it’s certainly more exciting to jump right into picking out flooring and colors, but it would be irresponsible to cover a cracked foundation with plush carpeting or hardwood.

How do I start?

Data Quality Management can seem overwhelming.  You may be thinking, “There is just so much data!” and “Where do I start?”  Start by looking at the situation at a high level and ask yourself these three important questions:

  1. Are data quality issues affecting our organization’s performance?
  2. If so, how can we measure and assign a value to the affect of these data quality issues, both financially and in terms of credibility?
  3. Does the value of having quality data outweigh the cost of a solution to address our data quality issues?

Then, assuming you have identified a valid need for a Data Quality Management initiative, start by clearly articulating the business reasons for undertaking DQM and outline the project.  These questions will get you started:

  • Business questions
    • Do you or your staff experience delays when requesting information, queries or ad hoc reports?
    • Does the board sometimes ask for information that you cannot easily provide without extensive manual modification?
    • Do multiple reports exist that report variations of the same information, but that sometimes have different numbers?
    • Are you confident that the quality of your data will enable you to make good decisions?
  • High level goals
    • What is the exact outcome you expect from the project?
    • How will use your data once it has been cleaned?
    • How will you define success?
  • Stakeholders
    • Who should be involved?
    • What is the time commitment?
  • Communication
    • How you will communicate about the project?
    • When and how often will these communications occur?
    • Who will receive them?
  • What is our budget?
    • Do we outsource some of the work?
    • What scan we automate?
  • What is NOT included?
    • If we are focused on current data sources then legacy data sources may not be included

Once you have answered these question, then define each business area you will include and then prioritize the list.  To make things easier for staff, take on one business area at a time, starting with the area that will make the highest impact for your organization.

What are the Data Quality Management Roles?

  • Program Manager/Project Leader
    • Since there can be multiple projects within the program, there may be multiple project leaders per program.
  • Organization Change Agent
    • Who is getting the organization culturally ready? This should be someone at a high enough level to influence change.
  • Business Analyst
    • Determines business requirements including data quality requirements
  • Data Analyst
    • Reflects business requirements in the data model and data acquisition requirements
  • Data Steward
    • This role is critical and should be considered a requirement for the success of the project
    • Manages data as a corporate asset
    • Gets business leaders to focus on data issues
    • Familiarity with all data sources

What’s next?

In our next two blogs in the DQM series we will discuss how to assess your data quality and how to tackle data cleansing.