Some people are looking to grow their use of metadata beyond an IT productivity project into something larger such as a data stewardship or data governance initiative. Some are just getting started. In either event, here are my top ten recommendations based on conversations with customers and some of the presentations and panel discussions I ran at Informatica World.
1. Show up, start small and execute. Honestly, showing up is half the battle. Pick a small project that you can be successful at and show some tangible results quickly. Then, grow from there.
2. Quantify everything. Be ready to quantify your results at the drop of a hat, literally. You never know when you will be asked to justify the existence of your metadata project. One financial services company at Informatica World showed how they saved over 95% on the analysis phase for planning data integration changes. Another health care company showed how they save a minimum of two weeks on every change analysis using search, data lineage and impact analysis. Have these metrics ready. Because the big benefit, good business decisions based on good data, is so many steps removed from IT it is often hard to quantify. Quantify what you can and keep looking to quantify other benefits. Example: What would be the cost of a bad investment decision because of bad data in a data warehouse?
3. Get executive sponsorship. Your project won’t succeed without the right executive sponsorship. Period. Don’t shortchange the importance of getting the right sponsorship from both the business and IT sides of the house. Several people have told me that this is what they did right after their first failure. This is also critical for getting other groups to contribute to the overall cause.
4. “No metadata project is likely to succeed without a data governance initiative”. With a data governance initiative, a company moves beyond metadata as an IT productivity tool and into use cases that have much broader business benefit across the organization. The most important thing is that you have a data governance council to set the overall direction and priorities for data-driven projects. They also need to design an overall framework for business users to collaborate with IT on these projects.
5. Pick a high-value target. Pick a specific problem and solve it before expanding into other new areas. For example: provide complete data lineage and full business terms and definitions around a new data warehouse or master data management (MDM) implementation. The important thing is to pick a project that has high value and high strategic importance to the overall business.
That first win is critical. (See also #2, Quantify everything)
6. Scope for success. I have seen a number of metadata / data governance projects fail because they tried to do it all at once. The projects quickly bogged down and then got cancelled when they failed to produce any meaningful results. Nobody has time or budgets for endless meetings that produce no tangible results for the business.
Often the failure comes from trying to resolve common business vocabulary across divergent business units. Start with a single project in a single business unit if possible and grow from there.
7. Get your business users involved. Great quote: “Data without business context has no value.” A table with a column name of NW_Net_Revenue has no particular meaning unless it is attached to:
- A business term
- A term definition
- A term owner
- Reference data: Does “NW” refer to a geographic region? What is in that region?
- Other documentation and comments.
Once you have this business context, you can link the business terms and definitions to the underlying technical metadata, creating a common lingua franca between business and IT that will improve communication and collaboration.
8. Use both the carrot and the stick. Think about incentives such as better access to business term definitions, contests for contribution, visibility, etc. It is also good to think about the stick. Having the backing of executive management and a strong data governance council can go a long way towards ensuring support and buy-in as well. The important thing is that it is important to use both approaches. All “stick” can be heavy-handed and result in only grudging support.
The challenge is to get the business community to want to participate. In almost all cases, it is not the full time job of the business side to provide this context. They all have full time jobs and are doing this “on the side.” The question is: What’s in it for them? You will have to find ways to incent them.
9. Tie metadata management to a business initiative. I picked this up after talking to two companies that had both tried and failed at data governance initiatives twice each. They both told me that their new approach was not a top-down data governance initiative, but to attach metadata management and data stewardship best practices to important new projects as they came through IT. This approach is much more pragmatic, scoped for success and likely to succeed. It is also more likely to show quick and measurable results.
10. Look for a data crisis … and be ready for when it inevitably happens. What if your management just won’t fund a metadata management / data stewardship initiative? A senior manager at a financial services company told me to “Just watch for a data crisis” and be ready with your proposal in hand. A crisis will happen. What matters is that you are ready with a proposal that shows how to prevent this type of crisis in the future.
These ten things won’t guarantee success but they will go a long way towards improving your odds of succeeding.