As I mentioned in my last post, we have produced an interesting conceptual model of IT as a business, from Business Motivation all the way through to Operations. For those of you who know I’m a SOA promoter: yes, I can represent a composite application in the repository.
More importantly, I can represent the reasons that a system exists, the metrics that demonstrate it’s value, the business rules that must be enforced, the business processes that take place, the people and organization hierarchy of the business, the programs and projects that drive change, the systems that deliver processing, security, and infrastructure services, and the operations that keep the lights on.
Microsoft IT is a big place (and apparently getting a little smaller, alas). So it is not possible for one team to “own” a model of this breadth without aligning with other sizable groups or divisions. There is no way to represent “one view of the truth” without creating a sense of shared ownership.
On the other hand, the motivations of each team can be different, and their perspectives on the model can be different as well. When creating a shared model, compromises must be made in order to get to a solution that actually works: rational, minimal, consistent, and aligned to the business.
With shared ownership, each compromise has to be carefully documented and signed off, because without that careful documentation, and the organizational will-power to refer to it, every compromise will be visited, and revisited, over and over ad nauseum. The noisy person and the politically well connected person will “earn” compromises in the model that challenge it’s quality and reduce its effectiveness.
On the other hand, as organizations change and adopt new ideas and well-organized principles, the model must change as well. We saw SOA add a different understanding of software applications and shared resources. We saw ITIL and MOF add a different approach to managing IT operations as services. Every year, something must change, even if it is just a little bit. Federated ownership is an excellent way to insure that those changes can be captured and incorporated.
But is federated ownership the only way? Of that, I’m not certain. If I look to the legal system, there is the body of law that has emerged through legislation, case law, and regulation that makes for a complex but workable system of rules and adjudications. There is no “owner” for all laws dealing with small businesses, for example.
In addition, if ownership is federated, what does that really mean? How does the central model stay consistent and useful when large subsections can change independently?
We have tried to create a central model that is minimal, that has only core objects and maintains only core relationships, so we can be immune from some obvious changes. On the other hand, this is an imperfect science. There is no way to be insulated from change.
Still mulling ownership structures. If anyone has any comments or advice, please share.
One thought on “On the federated ownership of a conceptual domain”
Good essay, good question. In my experience, federated ownership doesn’t work well, sort of like design-by-committee. On the other hand, I haven’t seen handing ownership over to a third party hasn’t working well either.
I think the answer lies in model governance. Leave ownership federated, leaving the owners with a stake in keeping the model current. The governance structure maintains the consistency and ensures that the model remains relevant.