Measuring Governance

I apologize. I said I would address this yesterday. We do have to get back to the guerrilla movement on the frontiers of the empire, but let's take a little time to look at measurement of governance.

First, let's agree that data governance is like any other governance except that it focuses on data. A governance program directed at process or at competency or whatever, would have the same characteristics? OK, I'll attempt a justification for that statement.

What do we ask of a data governance process? What are the objectives? By the way, I use the term process here in the sense of a set of activities that are ongoing and have a consistent purpose. The purpose of the data governance process is to:

Optimize the value of the data resource by insuring that the capture, storage, retrieval, destruction and use of the resource is done in accordance with established policy, procedure and standards.

Do you buy it? If not, I'd be pleased to discuss alternative purposes, but the remainder of this discussion is based on this purpose.

Based on the purpose of data governance then, several perspectives on measurement suggest themselves. The most obvious one is the QA (quality assurance) perspective. How are we doing at following established standards? It is tempting to count the number of standards, policies and procedures because counting is easy to do and there is a tendency among the governors to equate many laws with good government. Strangely enough, among the governed the emphasis is on the quality of the laws rather than their quantity. A small number of effective and easily understood standards may deliver more benefit than a larger number of over-specialized or esoteric ones.

The most effective measurement will be part of the standard or process itself, but some organizations may find it useful in getting governance going, to do retrospective analysis to see how well/consistently processes are being applied. Health care makes extensive use of the "chart review" to gather this kind of data retrospectively. Measurement intrinsic to the process or standard has the potential to be much more nuanced and useful than that done retrospectively simply because all of the context is available.

Clearly, though, the nature of the metric(s) is very much determined by the process or standard itself. For this reason, it makes no sense to discuss metrics or KPIs (key process indicators), a special kind of metric, without first establishing the process context.

Other perspectives might differntiate among standard, process, and policy or might measure in conjunction with the data life cycle, specific subject areas or specific usages.

One last point, should you be tempted to think in terms of measuring accountability.

Accountability in the absence of a standard is really approval.

No governance mechanism can exist for long based on approval. Each change in "leadership" will create massive turmoil as everyone seeks to reorient to a new approval model.