First, as discussed in an early post (Using Enterprise Architecture to Reduce the Federal Debt), changing organizational (governmental, and bureaucratic) culture is a slow painful process. Any process that requires cultural change will create an enormous amount of organizational process friction (particularly from transactional management--the gatekeeper--managers that know the rules, enforce the rules, live by the rules, and thus do not like changes to the rules). Additionally, it significantly limits the middle manager with a private agenda (the empire builder) to implement the private agenda. Generally, this takes a champion, like the CEO of the organization to participate any change; then the middle managers will slow roll it. Additionally, Subject Matter Experts may participate in this slow rolling, if they feel they will be out of a job, or if they feel they are in any other way threatened--Dr. Micheal Hammer speaks well to this overall topic.
Second, even if all of the cultural hurtles are overcome, the second reason for Enterprise Architecture losing any impact, is that Enterprise Architects feel the need for complete "as-is" architecture in the Asset and Enterprise Architecture Repository (AEAR) of the organization before they can start to use the process. Consequently, they ask for large sums of money and years to create the repository. The financial engineers of the organization and other looking for ways to cut the organization budget, look at this with jaundice eyes, as an appropriate cut, since there is no ROI apparent, only longer term VOI (Value On Investment).
However, there are ways to greatly reduce both the cost and time (to market) for creating the AEAR and ramping up the Enterprise Architecture process. More on this in future posts