So many tools, and so much metadata – how can a person possibly create a strategy that works for business and technical metadata? A friend of mine works for a company that basically has three different ETL tools. One is used for getting data out of the ERP system, one is used for getting data into the data warehouse, and one is only used for small projects. Are you confused yet? I sure was when my friend called and asked how to start a metadata strategy for her organization.

Here is what I told her:

1. Start with technical metadata
   1. Gather technical metadata requirements based on what that client needs to know about technical metadata (usage, lineage, impact analysis, job control, quality audits, etc.)
   2. Consider a way to bring the required metadata from the three ETL tools together in a dashboard or some reporting tool (please don't make three different reports for the three ETL repositories)
   3. Consider creating a separate repository for the technical metadata and your own ETL programs to extract the required technical metadata from the ETL repositories
2. Gather business metadata requirements
3. I would definitely create a data model to ensure completeness for this initiative

For business metadata, I rely on the data modeling tool repository. If I can share or transfer metadata from data modeling tool to ETL tool or to centralized repository, I do!

If you know of any other way to gather the technical metadata, please let us know!