I stumbled across a post by Brian Burke from Gartner called US Federal CIO Faces a Daunting Challenge, in which he points out that many government-wide enterprise architecture programs have failed because underestimated political, cultural, and financial challenges.

Though not stated in the post, one can then assume this is the challenge facing the US federal CIO, and the post then states the classical argument that they are given the responsibility to optimise ... without the authority to get it done.

Most anyone who's had to implement change in a big organisation, have at some point wished they had the authority to just command what needed getting done, and sure, given a stick big enough, change can be forced upon an organisation - but the end result is not very good and it causes as lot of dissatisfaction. If empowerment was the universal solution to the challenges facing the US CIO, then it would be an easy fix.

True, if you harmonise enterprise architecture in a loosely bound group of organisations or even just within a single organisation, this impact will affect some more than others, but that does not imply that there must be winners and losers. Taking a note from economics, a Pareto efficient solution can be sought so that no one party will worse of after the change.

This can also be viewed as addressing the political challenges of a project: Say to someone that you're going to change their life, and that they're going to be worse off afterwards, and they won't like it. On the other hand, if they're going to be better of afterwards, they are much more likely to agree.

In a Pareto efficient solution, the end result only needs to be on par, but because people are predictably irrational, it often needs to be better. Specifically, people tend to value higher what they have, which is also referred to as the The High Price of Ownership or endowment effect, and so to make people agree to an alternate solution, they have to value that solution as high as the inflated value of the solution they currently have.

So, instead of forcing an organisation to accept a specific solution, the chances that a solution will succeed, can be greatly improved by understanding the values of the recipient, and making sure that the recipient is not any worse off.

I spend most of my time building federated security solutions, and Brian is the unofficial author of the definition federated architecture, so I would have en enjoyed a perspective from him on how instead to pull together the values of technical infrastructures in different organisations using federated architectures - rather than dreaming of unity. But that's a topic for another post.



blog comments powered by Disqus