The word governance is often uttered with much gravitas by a senior consultant towards the end of an ‘engagement’. Governance usually implies that something more has to be done, and paid for, before whatever it was they were brought in to do can deliver its full promise. I could talk about Governance in Testing as that’s a part of what we do for a living, but instead I’ll discuss Governance in a different aspect of IT, Corporate Technical Architecture (CTA).
A CTA is essentially a process of selecting and managing a set of standard IT components for an Enterprise. The idea being that all new implementation projects will use the standard components and: save re-inventing wheels, reduce support and operations costs, and facilitate any future integration every large organisation should have one!
There are standard ‘methods’ to produce a CTA them and recommended ‘Governance’ process to maintain it. However, the effectiveness of a CTA is usually determined by company politics. The most effective CTA I ever saw was in an IT organisation where CIO was an outright dictator. They personally made all the technical decisions and their minions implemented them, or left if they failed to. The standardisation and simplification achieved was outstanding. On the other hand there are large numbers of CTA which are either almost content free or run large numbers of approved exceptions. (A CTA expert would claim that this is OK as even having the dialogue is of value).
To the point then, Governance at its core is about stopping another part of your organisation doing what it wants to. Before you embark on anything that will ultimately need Governance to make it work, do the political mind games of; where in the organisation’s hierarchy would the enforcement have to happen and how bad would the situation have to be before it could be made to stick? That can tell you more about the process you need to build than almost anything else.
One last point, Governance is about being able to stop things but being aware of things that you can’t stop is of value, for example, knowing a plane is about to crash will not prevent it but at least you can adopt the brace position.