Assessing Data Quality Process Breakpoints

Fixing bad data is more than just a technological challenge

Data quality goes beyond name and address cleansing routines or hopes to correct data at the entry point. It requires rigor around requirements, clear organizational responsibilities, and a cultural awareness that mandates continuous data improvement over time.

Baseline’s Data Quality Audit uses a six-step process to diagnose your bad data, illustrate its impact on the business, and prescribe ways to fix it.

» Your Challenges
» The Problem
» The Baseline Approach
» Your Value
» Why Baseline

Your Challenges

• Core business initiatives imperiled by bad data
• Formalizing the multi-step data quality process
• A structured and sustainable approach to data definitions and rules
• Readiness for a data quality center of excellence
• Engaging the business in data ownership
• Defining job functions, roles, and organizational structure for data quality
• Selecting data-cleansing technologies

» Back to top of page

The Problem

The hidden costs of bad data

When data quality hits the radar, it’s usually at the business’s expense. The customer notices a billing error on an invoice—for the third month in a row. Or a finance executive knows that the top-line financial reports are wrong, but he can’t put his finger on why. Or the regulatory agencies come calling, and there’s contradictory data.

Indeed, poor data quality has been the unfortunate discovery of many companies in the throes of deploying new enterprise systems, releasing new reports, or launching “single view of the customer” initiatives. These days, business users are increasingly aware of “bad data” and are raising data quality improvement as a strategic business issue.

Does any of this sound familiar? If so, think “data quality improvement”.

• Contradictory reports have become an executive-level problem.
• Lines of business have different definitions and rules for common data.
• Programs that share information across lines of business, like business intelligence (BI) or enterprise resource planning (ERP), have incited debate about data definition and ownership.
• Technical issues that affect data quality—such as slowly changing dimensions, lack of referential integrity checking, or competing ETL programs—lack the resources or consensus necessary to fix them.
• There is minimal understanding of key source system data in the enterprise.
• A high-profile IT project has failed, been delayed, or been “de-scoped” due to poor data quality.
• No one owns the problem of fixing data defects at the source system level.

» Back to top of page

The Baseline Approach

Instilling quality, a data slice at a time, across six discrete process steps

Baseline’s Data Quality Audit is a bottom-up activity. We begin by diagnosing and illustrating the problem of bad data before prescribing ways to fix it. This includes evaluating metadata, which often gets overlooked as a data quality determinant.

We examine how your company manages data across the six primary steps in Baseline’s data quality process and we identify the break points. The Data Quality Audit examines:

• Defining business and data requirements.
• Locating data sources for data acquisition.
• Profiling data as an ongoing, automated activity.
• Standardizing, validating, and correcting data, including existing ETL programs and the business rules they use.
• Match and merge functions, including data enrichment and consolidation.
• Deploying formatted data to the target system and monitoring its accuracy.

Baseline then customizes an action plan that your company can use to implement data quality at each point in this process. We recommend organizational responsibilities and cultural changes. We help you evaluate and improve the use of data quality and data profiling tools. We look at the role of systems architectures, like service oriented architecture (SOA), in provisioning data quality as a service. And we define data quality metrics and success factors for continuous tracking and reporting.

Finally, the Baseline approach endorses starting with a small data subset—a “slice in time approach”—to prove and tune the data correction process to meet your company’s unique needs.

Data Quality Process

» Back to top of page

Your Value

Proof points to trigger investment and change leadership for sustained data quality

When Baseline presents our findings to your business and IT management team, they will not only understand the current impact of bad data, but also its future costs and the benefits of investing in data quality improvements.

From a bottom-up perspective, the Data Quality Audit illustrates “what bad data looks like”. From a top-down perspective, the audit identifies the key business impacts associated with poor data. Baseline’s Data Quality Audit can propel your company away from “bad data” corporate anecdotes and toward a structured and consistent approach to finding, fixing, and maintaining data for both operational and analytical purposes.

» Back to top of page

Why Baseline

Data quality rigor—a regular by-product of every engagement

Baseline is an acknowledged specialist in data integration and management. Driving rigor around the data quality process is “what we do”. In fact, data quality is a regular by-product of all our work in business intelligence, data warehousing, and data integration.

Baseline understands automation as a key factor in delivering data quality. We have significant experience working with all of the leading data quality tools—including DataFlux, Business Objects/Firstlogic, Trillium, IBM/Ascential, Informatica/Similarity, and Group 1 Software. We bring an in-depth understanding of their offerings, as well as their strengths and weaknesses, to the Data Quality Audit. At the same time we maintain our position of vendor neutrality and do not resell hardware or software.

Baseline consultants work with both IT and the business to determine how to leverage existing processes and business rules. We help companies formalize a data quality program or center of excellence that can serve the enterprise for the long-term.

Data Quality’s Role in Master Data Management

» Back to top of page