Chapter 4. Implementing a Metadata System - The Key Points of This Chapter

• Metadata systems are built around existing data systems and, ideally, the organization's vision for future data use and management. Introducing a metadata system is a complex project that requires planning efforts comparable to any other major initiative. Unlike other technology initiatives, however, an organization's stakeholders in most cases will not have asked for a metadata system because they may not yet appreciate its potential for helping them better understand and use data.
• Issues specific to implementing a metadata system include
  • building capacity given that most organizations do not have sufficient metadata expertise within staff ranks;
  • identifying metadata needs even if stakeholders are not familiar with the concept and potential of a metadata system;
  • justifying metadata through strict cost-benefit analysis given that many of the benefits of a metadata system cannot be quantified by traditional market pricing;
  • determining whether to choose an off-the-shelf product given that a metadata system must be customized to meet each organization's unique data needs and user requirements;
  • determining whether to pursue a centralized, federated, or distributed system architecture;
  • establishing a clear understanding of the organization's existing data—what data exist, their format, their location, and their quality—before expecting a metadata system (which is dependent on those data) to provide useful information; and
  • training stakeholders to understand metadata as a concept before they use a metadata system as a tool for improving the management and interpretation of the organization's data.