Expanding Data Governance: Where to Start and How to Succeed

Data Governance programs often fail - not because of technology, but because they are treated as a set of MDM-related systems rather than a strategic process. Whether initiating a new data program, formalizing one that already exists, or consolidating multiple decentralized governance initiatives, it’s difficult to know where to start. The first thing to think about is program goals. Those goals need to be distilled into concrete, value-driving work packets that then become the roadmap.

Program Goals

The push to formalize MDM often starts with top-level aims such as:

·      We’d like to improve our data quality.

·      People don’t know the single source of truth.

·      We aren’t using data as a strategic asset.

These are strategic aspirations that need to be translated into actionable, measurable goals granular enough to guide a program. Goals need to be drilled down multiple levels to the place where sets of specific data elements are identified that can impact a business outcome. Each of these sets supports use cases for data. A use case might be a report that isn’t trusted, it might be a decision of what product to show a user on a mobile app, or it might be data related to parts supporting a specific step in a manufacturing process.

Identify a portfolio of a dozen or more use cases, but focus on detailing a core set of 3-5. These should be impactful, deliverable, and serve as a blueprint for future initiatives. A CFO thinks in terms of quarterly results, and the program needs to show value fast. These use cases are not just the roadmap of the program: they are the roadmap to value.

Picking Where to Start

During the planning stage of identifying and detailing out use cases, most of the information needed to decide where to start will become known. The shortlist of 3-5 opportunities should share certain qualities: clear value in terms of supporting better decision-making and action, a small enough set of data elements that within a maximum of 3-6 months they can be under active curation, and stakeholders who are engaged.

A few additional guidelines to consider:

1) Engaged and enthusiastic are not the same thing, and enthusiasm without availability will kill momentum. Because the business stakeholders will be critical to program success, make certain ahead of time that the stakeholders will be available for the amount of time they are needed. Use behavior during the planning stage to gauge this: stakeholders who were tough to engage or had limited availability during planning will be equally or more so during execution.  

2) Size and complexity matter. There will definitely be “lessons learned” from implementing the first use case, however assuming it is mostly successful it should be useable as a blueprint for future use cases. It should not be so small or simple that it doesn’t serve as that blueprint. It also should not be so large or complex that it extends past the 3-6 month time horizon or gets tangled in its own complexity.

Apply All Aspects of the Program to the Use Case

Multiple systems and processes support Data Governance. The organization may be expecting some of these processes to be done manually, supported by a spreadsheet, or may have purchased one or more of the many excellent systems that exist for things like data discovery or data catalogue. In the latter case, there is a temptation to want to maximize the use of the new tool … for example, do data lineage for all use cases. Avoid this temptation. A tool is not a program, and the goal of a use case is not just learning how to use a system. The objective is to deliver the full business value, using only the necessary tools to support the long-term strategy. Overloading the first use case with every possible tool or feature risks slowing down the process and distracting focus. Treat the use case as a pilot of the program, not the tools.

Communicate Value

The senior stakeholders who initiated this journey should receive communication about its success, with updates to the roadmap and lessons learned. Ideally they should be hearing it from not only data program leadership but from the business stakeholders who are engaged with and benefiting from the program. “We increased revenue in part because with better quality data we were able to do xyz: the data program is a business accelerator.”

Conclusion

Consider this “checklist” when rolling out or formalizing a Data Governance program:

·      Start with business value: Specific use cases that improve decision-making or operations.

·      Engage the right stakeholders: Engaged and committed partners who make themselves available.

·      Favor manageable complexity: Interesting enough to serve as a model and small enough to execute in a quick timeframe.

·      Implement the full process for the first use case: Address all aspects of governance (quality rules, glossary, lineage) rather than applying partial solutions broadly.

·      Communicate success in business terms: Show that data impacts top- and bottom-line results in a direct, understandable way.

Avoid the common pitfalls of vague goals (“better data”), a systems-first mentality, or implementing partial solutions broadly. Targeted effort, fully executed: scale with confidence!