A tough question. I thought it was some sort of common good – lots of people own bits of it, and all those bits come together to provide the services we know and love.

But that wasn’t the question being asked. The real question was: “Who in this company is accountable for what we do on the web?”

Have you ever noticed how questions like that come up mostly when things are going wrong? If the project is ahead of schedule and winning plaudits from all stakeholders, then people rarely start looking for owners. It’s only when we run into problems that we think about governance.

We all know this is wrong. For a start, while we’re looking for an owner we’re doing little to actually solve the problem. And if the owner had been defined up front, then they might have acted to prevent the problem occurring in the first place. Yet we persist in leaving ownership and accountability fuzzy. Why is this?

I see three common reasons:

  • Firstly, we ask simplistic questions. No one person owns the web. Not in its entirety and, as a general rule, not within the confines of any company. Ownership of the web, or the customer, or the SOA, is complex, because these are complex entities. Seeking a simplistic answer quickly becomes intractable for a host of reasons; political, technical, historical, and intellectual. On the other hand, if we set out to understand accountability for a complex network of interacting factors, then we have some hope of unravelling it.
  • Secondly, defining clear governance is hard work. You need to develop a clear view of what assets must be governed. You need to identify all the people who have a stake in each asset. You need to resolve boundary disputes. This work isn’t sexy. Many people would rather just get on with the fun stuff. If you think crisis management is fun, then that’s a good approach.
  • Finally, many people have an interest in keeping things fuzzy. If we don’t know who the owner is, we can’t blame them when things go wrong. If the boundaries are unclear, then people can grab more territory for themselves.

Clear governance is possible. I’ve seen places where the organisational assets were clearly mapped out, typically into some sort of well-structured enterprise architecture. These places then recognised that different ownership models could apply to different assets – some were owned by a single person; some were jointly owned by a duopoly; some were federated and managed by some sort of governance board. They didn’t try to fit every asset into a single, simplistic model. I commend this approach to you.