The marriage of master data management and data governance was probably not what singer/songwriter Jack Johnson had in mind when he composed “Better Together.” But MDM and data governance are indeed a power couple whose union is equal to more than the sum of its parts.
As data sources and volumes continue to mushroom, so does the mandate to understand and implement strategic practices to extract the full value of the data, not just manage it. Making business data consistent and accessible across the enterprise is a critical first step in taming and putting big data to use.
An organization’s ability to identify, access, integrate and synchronize data from multiple internal and external sources and then effectively govern that data over time using processes and policies is crucial. Collectively, the focused effort to manage master data coupled with an organization’s ability to leverage sources of big data, especially social networks, will result in optimized business processes and measurable business performance improvements.
While there are different philosophies on the best ways to achieve MDM, one movement gaining traction focuses on the need for an evaluation of the entire data ecosystem, including processes, resources and deliverables. And while some still debate the primacy of MDM or data governance, the focus now needs to shift to unifying MDM and data governance to overcome both organizational and technical issues. The next evolution of managing and governing data isn’t about deciding which comes first. Rather, the two related but distinct practices need to merge into a discipline called master data governance. This point of view is supported by independent industry experts, such as Aaron Zornes, chief research officer for The MDM Institute.
“Enterprise-level governance that spans both data and process is increasingly a key requirement put forth by IT executive management. While MDM purports to span the entire master data lifecycle – including creation, cleansing, harmonizing and archiving – business process management claims to dominate the same for the business process lifecycle. Such dogma makes it extremely difficult to execute either MDM or BPM to their full potential,” said Zornes. “We believe that the solution is for master data governance to unify these two worlds to overcome both organizational and technical issues.”
Before we examine the combined practice of master data governance, let’s review its components.
Master Data Management and Data Governance
Any discussion involving the topics of data management and governance requires clarity around the types of data. For purposes of this discussion, we’ll focus on three different types of organizational data: master, transactional and metadata.
Master data is the heart of business optimization and refers to core business data. This can include look-up data, such as standardized lists (e.g., product type, product category, unit of measure). It is data that changes infrequently and requires simple data governance and maintenance. Also commonly found in this data group are master attributes and key data. These are subsets of master data that define business objects such as products, customers, distributors and locations. Finally, master relationship data details data relationships among master data entities, such as customer, to product relationships and customer and product hierarchies. Once recorded, master relationship data constantly changes and requires ongoing governance and maintenance; this can be complex and a challenge for many organizations.
Transactional data describes discrete events and, once recorded, never changes, requiring no maintenance. They occur as discrete points in time and reference master data. For example, a record of business events such as orders, forecasts, inventory and sales history are classic examples of transactional data. Metadata refers to information about the data itself such as data type, definition and constraints.
Data governance can be defined as the practice of organizing and implementing principles, policies, procedures and standards for the effective use of data. Governance is essential as data volume grows if an organization expects any of the data to be useful. Governance, organizational and process changes are common barriers found in the majority of organizations despite significant progress in technology. As a result, companies of all sizes struggle to ensure a single version of the truth for their products or services and the associated data across their organizations in heterogeneous IT environments. The source systems are not in sync, are overloaded with stale data and simply cannot be accountable to deliver simple analytics such as “sales by customer across business lines.”
Master Data Governance
Achieving master data governance requires organizational commitment in the following areas: enterprise data management, education and alignment of staff resources, and the establishment of a business-oriented enterprise governance of data.
Organizations should take a centralized approach with a master data governance process that supports the global identification and linking of “master” business entities and the relationships between them across the enterprise. This can include customer, product, geographic and reference data. They need to provision a complete data governance solution including workflow, business rules, user interfaces and data quality. A common multidomain data model should be used, and the process should create and manage a central repository of master data with real-time integration and synchronization of a single view of master data. Ideally, it should provide ongoing master data stewardship and governance capabilities through monitoring, corrective action, workflow and reporting.
Common core challenges with master data governance include the lack of access to a single, accurate view, with multiple sources not in sync across the enterprise. It is also common to see multiple business hierarchies using the same data and complex role-based workflows. Another pitfall involves placing too much focus on where the data is stored. Traditionally, many organizations have opted to create an operational data store that could be updated in real time or done offline. If structured properly, a data warehouse can act like an ODS, but an ODS can never perform as a data warehouse. The focus needs to be on the long-term needs of master data across domains in the enterprise.
Focus and Align Staff Toward Delivering Business Value
As the data volume grows, so must the skill set of staff. Big data scientists, enterprise architects, data stewards and line of business managers are now tasked with solving a very real and critical business problem: enterprise governance of data. The staff must be aligned around creating a centralized approach to master data governance. This alignment can be accomplished by creating a thoughtful dialogue within the organization that defines a single view of your master data.
Once the core functions and desired outcomes are understood, start developing a plan that is measurable, repeatable and leverages the technology infrastructure. Increase the chance of success by starting with a funded project that can be tied to a business metric. Remember that this is a journey and should be positioned accordingly and measured through phased deliverables. Finally, tell the story – through a roadmap illustrating the domains servicing various business functions with regular updates. Communication is vital to long-term success.
Five Key Takeaways
- Unifying MDM and data governance is essential to overcome both organizational and technical issues and addresses the need for evaluation of the entire data ecosystem, including processes, resources and deliverables.
- Master data governance is directly linked to business performance, and organizations that commit resources to data/business improvements and management will always outperform competitors who don’t.
- Instead of blaming it on an organization’s political or cultural climate, the misalignment or mismanagement of data should be viewed as an opportunity to create real business value.
- The importance of generating a single view of customers, products, locations and relationships cannot be stressed enough.
- Organizations can optimize business processes across functions, leverage their existing technology investment and build an integrated data workflow.