By Barbara Latulippe — Sr. Director, Office of Architecture and Innovation
Like most organizations these days, your company has probably realized the crucial role Data Quality plays in driving operational efficiencies and leveraging business and predictive analytics.
Be aware, however, that managing master data and driving data quality is a long-term journey that is best approached in phases based on your needs. If your organization is just starting out, it usually takes three to five years to get to a matured/optimized model, so I encourage you to take pragmatic steps and focus on your important business priorities. It’s not a one-size-fits-all endeavor.
At EMC, we are in the midst of evolving from a managed to an optimized Master Data Management (MDM) and data quality model. Over the past several years, we established and implemented that model for our customer data domain and are in the process of applying that same approach to additional domains, such as contacts and vendors.
Along the way, we have learned some lessons that might help your organization with its data quality journey. These lessons include:
Define master data management for your business
To begin with, let’s define master data management. According to information technology research and advisory firm
Gartner, Inc., MDM is a technology-enabled discipline in which business and IT work together to ensure the uniformity, accuracy, stewardship, semantic consistency and accountability of the enterprise’s official shared master data assets. MDM is made up of data quality, data governance, architecture as well as an MDM framework itself.
In EMC’s journey to improve our information management, we are focused on the following competencies of the MDM framework:
- Ensuring that the vision and mission of information management is well understood
- Optimizing architecture and how data integrates with it—this includes where is information consumed, how is it consumed, who’s consuming it?
- Information quality
- Metadata management
- Establishing the organization model to support stewardship and governance
- Arriving at a single source of truth
I will tackle the benefits and challenges of each of these areas in more detail as this blog series continues, but first let me give you an overview of EMC’s initial approach to our MDM journey.
EMC IT initiated the push to advance data governance and the vision and implementation of an MDM framework in late 2011. In addition to the overall need for ensuring data quality, MDM was a key factor for the success of several major IT projects, the largest one being our PROPEL project, an initiative to replace the company’s decade-old ERP with a new SAP-based system.
Project PROPEL leveraged the governance model initially to help us execute a successful data migration, achieve our go-live milestone and have an effective change control process ensuring the integrity of our solution post go-live. It was key that we have agile decision-making, change management and knowledge management policies and processes in place to safeguard our investment in our ERP.
As with any good data governance framework, it was crucial to focus on the business challenges and to quantify the cost of poor data quality to business metrics. After all, Gartner estimated in a 2011 report titled, Measuring the Business Value of Data Quality, that poor data quality effects overall labor productivity by as much as 20 percent.
We chose to begin our MDM process by focusing on our customer data. When we began, EMC had several pockets of data governance throughout the organization, but the data governance framework sought to unify those practices into a more structured, focused working team.
Understand your business drivers and pain points
To understand where to start with your MDM framework, you need to understand your organization’s business drivers, key projects, strategies and business pain points. To do this, I first met with key stakeholders from each business group.
We then established a Data Governance Council made up of 20 employees at the operational/tactical level of the business—subject matter experts who were close to the data. The council was instrumental in understanding how the data was created and consumed and helped us create the vision for business intelligence (BI) and predictive analytics to achieve competitive advantage.
Business users began logging their data issues—such as missing or poor data or a lack of standards and policies for data use—with the governance council. This quickly uncovered the need for a single source of truth which led to the company’s vision for a single customer hub.
Create a roadmap to MDM
Based on that input, we started to develop a roadmap of where we needed to go and who would be early adopters of the framework.
Other steps to consider in getting started include:
- Secure strong executive support
- Understand your current maturity level and develop implementation steps to drive it to a more managed and mature system
- Leverage documentation already available from groups who are “in flight” in establishing a data quality process and fill in the gaps
- Start small and grow participation
The MDM framework helps guide you on issue management, setting up a network of data stewards in the organization, creating the overall architecture and truly understanding how data is created and consumed in your company and tying that back to key business processes that are not optimized as a result of poor data quality.
When we began this effort, there was very little awareness at EMC about the business impacts of data quality. We ranked at the very beginning of a 1-to-5 data management maturity rating system established by Gartner.
In the past two years, we have strengthened our data management ranking to a level 2.7, and are looking to achieve an optimized model of 5.0. We have shifted to a top-down approach to management with more executives buying into data quality. We have improved our data quality at the individual domain levels, such as customer, vendor and material records. Data stakeholders have a much better understanding of how they impact each other when creating or consuming data. And we have data quality metrics in place to help us monitor compliance and identify new areas to drive issue resolution.
In the process, we have established a single source of truth for customer data and we in the process of establishing a universal customer identification so each customer can be uniquely identified in any process or application throughout the company. We expect this Universal ID to be in place by the end of the year. We will then work to extend that success out to other data categories, including vendor, contracts, parts and products.
From there, we will continue striving to improve out MDM processes to reach level 5—which is a very service-oriented and optimized model where data quality is built in to our processes and the business views data as an enabling asset which is managed and mined for insights.
Stay tuned for my next blog which will focus on the organizational model for data stewardship and governance.