Data governance definitions

Mixed messages

I am getting mixed messages about data governance. IBM recently published the results of a survey it had conducted into the use of data governance, which it conducted in conjunction with the NCC. Out of 141 respondents, from companies of all sizes, only seven per cent reported that data governance was neither implemented nor on the planning horizon. Of the remaining 93 per cent more than half had some sort of (at least initial) process in place. Now, admittedly this was a self-selecting set of respondents but this does not fit with my experience.

As a counterexample, I asked SAS (and bear in mind that SAS owns Dataflux, the data quality vendor) last summer about demand for data governance across Europe. "None" was the answer that I was unequivocally given. Now it could be that I was speaking to the wrong person but bearing in mind that this was a special analyst conference put on by SAS EMEA then this is unlikely: they don't put people in front of analysts who don't know what they are talking about.

Aside: SAS EMEA is now defunct (almost immediately following the retirement of Art Monk-to which there is [ahem!] obviously no link) and judging from the recent surge in SAS contacts to my Linked-In contact list then there are probably some good technical and marketing people looking for new opportunities.

To return to the main point, either something has radically changed in the last six months or we are not talking about the same things. I suspect it is mainly the latter, which means we have a definitional issue. It may also be that the UK is ahead of the rest of Europe in this area though we are, I think, some way behind the States.

In so far as definitions are concerned: first there is compliance versus governance and secondly there is data governance versus data quality.

As far as I am concerned, data governance is more than either compliance or data quality. In essence it is about ensuring that data is available, secure, audited, traceable and fit for purpose. The need for data quality procedures is implicit in that definition, while compliance is a natural consequence of it. Note too, that I do not believe that compliance is simply about the fulfilment of whatever legislation or regulatory framework is in place. It is in fact about meeting best practices and while these are sometimes imposed externally, they may also be established by the business. Indeed, it is reasonable to assert that the only reason why we have Sarbanes-Oxley, Basel II, MiFID and the rest is because business has failed to keep its house in order: external legislation, in this regard, is simply the imposition of best practices that businesses have failed to implement for themselves.

Data governance then, or indeed any form of governance, is about establishing the policies and procedures that the company should adhere to, while compliance is the process of ensuring that those best practices are followed. A company "doing" data governance therefore has a body established to define those policies and procedures, with enough power to ensure that they get enacted, and mechanisms in place to ensure that these are fulfilled: data quality software, dashboards and so forth are merely tools (if important ones) that enable this. My suspicion is that far fewer than 93 per cent of organisations are actively engaged in data governance in this sense.

Copyright © 2007, IT-Analysis.com

Sponsored: 10 ways wire data helps conquer IT complexity