Where, oh where did the metadata go? I still don't see many solutions that are easy to maintain and include all of the metadata required to show ROI on a project. Most people are still talking about metadata but not implementing complete solutions. What I would like to share with you is how to create what I call a metadata roadmap.
The dictionary definition of a roadmap is a visual of the highway network showing names, locations, route numbers, types of road and mileage between points. A metadata roadmap can be defined as navigational information showing names, locations and other pertinent information about the data and processes required to run the corporation. Integrated and easy-to-use metadata can answer the following type of questions:
- Who is using the data?
- What data elements are used?
- When is the data used?
- What is the quality of the data?
- What is the definition of an entity or attribute?
- What processes use a specific object?
To create a metadata roadmap, include the following steps:
- Gather and define business and technical metadata requirements for the corporation.
- Create a corporate metadata strategy based on reviewing multiple solutions and approaches to metadata management, understanding tool interactions and the metadata each tool stores and modeling the required metadata.
- Set next steps for a phased-implementation approach.
Gathering requirements for metadata will require some education for the business workers. Explain exactly what metadata is and is not. Start the session with general knowledge transfer on what metadata can mean to a corporation. Include scenarios where metadata can make a difference and is equal to productivity improvements, such as:
- Reduction in data redundancy;
- Program reuse to reduce redundant development work;
- Understanding how information is used in the corporation;
- Impact analysis to categorize simplistic versus complex changes;
- Better use of data for the corporation;
- Knowledge transfer, information sharing and navigation of the corporate data assets;
- Inventory of all corporate data assets; and
- Identification of corporate discrepancies and overlap.
A corporate metadata strategy is the policy and set of procedures used to manage the processes and the meaning of the data within a corporation. The metadata strategy is part of the governance and stewardship program and should be adhered to during any development project. Every project should have a step (or two) for metadata in the project plan, ensuring that:
- The business definitions are in the data model and propagated correctly into the flow of metadata for the company.
- Business definitions are incorporated into the front-end applications for the business user's easy understanding.
- Technical metadata is incorporated for future use, such as impact analysis and module reuse.
Metadata should include source system information, entity and attribute definitions, system usage information and an understanding of what information is fed from BI to other systems. On the other hand, technical metadata should include an understanding of the interactions between all extract, transform and load (ETL), reporting, analysis, database management, data modeling, data quality, data profiling, usage and monitoring products.
Metadata is found in every tool used in the enterprise! Database management systems store the source systems as well as the BI applications. The data dictionaries are easily read by the data modeling product set. Data modeling tools and target system data models have source system entities, attributes, relationships and definitions. Data profiling and quality products have repositories with metadata about the data analysis and quality anomalies of the data. Enterprise application integration or enterprise information integration products have repositories that show the relationships between multiple data stores in the corporation and how the data is used. ETL tools have process and program/job information.
The next step in creating the metadata roadmap is to review current market solutions for metadata, which are:
- Purchase a repository,
- Purchase a specialized metadata management product,
- Build your own repository, and
- Use a hybrid approach.
In some instances, using the ETL tool (hybrid approach) as a metadata hub may be the first phase of integrating and understating corporate metadata. The metadata model can be created from scratch, or many industry sources have models to use as a starting point or for reference. Keep the model simple and flexible, then each phase can be added easily.
To avoid mistakes while creating the metadata roadmap:
- Gather metadata requirements first,
- Map out different phases of implementation,
- Don't take on too much in the first phase,
- Understand and capture useful metadata from tools,
- Automate the metadata flow as much as possible,
- Collect and disseminate metadata as part of the development methodology,
- Include business and technical users in the process, and
- Don't create a metadata repository from scratch without resources for upkeep.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access