LESSON - Data Warehouse Governance: Ensuring Lasting Business Intelligence Value
By John Williams, Vice President of Technology, Collaborative Consulting
Even when constructed efficiently, data warehouses are expensive. As a consequence, companies that invest in them must ensure they receive the best possible return on their investment. A comprehensive data warehouse governance program ensures that business intelligence assets are expended for business benefit—both directly and indirectly—and that optimal value is achieved from the data warehousing environment.
Continuous, vigilant governance is critical. Without it, a data warehouse initiative will probably fail, even if it is a tactical success.
Key Elements of a Governance Structure
To execute a governance program, an organization must combine three key elements: sponsorship, organization, and process.
Ensure executive sponsorship.
Successful data warehousing programs feature thorough executive sponsorship, i.e., solid, enthusiastic senior management backing. Without that, governance programs, like most programs, fall flat. Senior management must support the program and (most important) provide funding and access to resources for data initiatives.
Tip: The champion must sit at least a level above individual constituencies and lines of business.
Structure an oversight organization. Establishing the appropriate team is another key governance component. Start with a governance board comprising senior business and technology contributors. They will ensure that the right people provide direction and have a vested interest in the success of data initiatives.
Data owners, data stewards, and data beneficiaries are also critical to a data warehouse organization. Owners include groups that provide data to benefit the organization; they own the content and the corresponding definition of quality. Stewards manage data on behalf of the organization. They execute processes that support the organization’s SLAs. Beneficiaries receive value by using information. The governance model must consider brokering dialogue among these constituencies. Then, data collection, management, and use can achieve optimal value.
Tip: The governance board should feature an executive sponsor, as well as business-unit and information-technology decision makers.
Establish lasting processes.
Once the proper sponsorship and organization are in place, fundamental processes can also be established. These focus primarily on alignment, prioritization, funding allocation, measurement, arbitration, and program management. The directed execution of these ongoing processes allows the data warehouse to provide clear near-term and optimal lasting value. Most data warehouses that fall short of expectations lack the process discipline to ensure success beyond the first few releases. This shortcoming often reflects an insufficient focus on and discipline in process management.
Tip: Processes must focus on long-term success, not just near-term tactical goals.
Imperatives, initiatives, and requirements.
A governance model must, first and foremost, consider the needs of the business—e.g., profitability growth, compliance, and operational efficiency improvements—as its primary driver. Framing the business problem first, apart from specific data-centric activities, provides clarity and helps teams focus in proper areas. This can be achieved by defining imperatives, initiatives, and requirements.
Imperatives represent the organization’s business requirements and objectives. Initiatives are data warehousing activities that help achieve imperatives. Establishing a direct correlation between both and leading efforts accordingly clarifies a data warehouse’s value proposition. The priority of business imperatives, and the effort required to accomplish them, drives the initiative-based program, and therefore keeps data warehouse efforts in lockstep with business needs.
Requirements are specific details of what is needed to support business imperatives. Initiatives group requirements into executable projects, which are managed and directed by the governance structure.
Maintain High Performance
Data is extremely valuable for companies able to exploit it—despite the time, expense, and complexity involved. However, when one considers all the effort required to establish a highly functional, robust business intelligence system, constant, diligent, and vigilant governance only makes sense. Otherwise, entirely preventable problems may arise, hindering performance and reducing value.