This is a hot topic amongst my colleagues. Federated governance take many different forms and the question remains, how much is enough to maintain the value of Enterprise Architecture? The landscape in a federated world is affected by the maturity of the Architecture practice within the Corporation. This is true of both centralized and federated corporations, however it is especially critical in in conjunction with a federated scenario. Federation is difficult for an mature EA practice because there needs to be a large organizational shift to move to federation and once there, finding the right mix of EA centralized control begs the question, do we even need a centralized governance function? On the other hand an organization that has an immature EA practice would most likely have trouble aligning it’s governance structure directly because each federated portion has been trained to operate independently. The operations are separate and distinct. According to Forrester [1], the EA practice number one driver is cost savings. This indicates that should inception of a practice come to fruition, cost savings through consolidation is a main driver of the practice. Federation at this point of maturity finds itself in conflict with the consolidation efforts. It would be suggested that an organization mature into federation. This would provide the continuity and communication channels organically rather than authoritatively.
As it would follow, EA governance needs to be grounded centrally in order to apply to the organizational goals. Alex Cullen of Forrester points out that only 35% of EA practices are centralized, with 26% federated and 22% meeting a hybrid approach. The relative high federation and hybrid (Total of 48%) EA practices call for a high degree of standardization [1]s. In 2015, firms reported a high degree of system standardization across business units, but much less in the business process and analytics spaces spanning the organization. While gains are found in sharing infrastructure, the processes that support them are still unaligned providing an opportunity for standardization within the process space, ultimately a catalyst for deepening the integration of systems. This is particularly difficult to achieve because of specialized and deeply ingrained operational nature of processes. EA governance is dependent on corporate governance and IT governance to support the function. Without the two working in concert, EA governance would be virtually unmanageable. Corporate and IT governance provides the overarching policies and standards that solidify the backbone of EA. Enterprise Architecture provides a lighter governance function that offers guidance on patterns, frameworks and communications between the business units.
Implementation of a governance practice is very difficult and no two look alike. [2] Though the federated model, EA governance finds itself split between a light model, where in practice, EA has no teeth and is only able to deliver the message “I told you so” and conversely, being overrun with requests for oversight. The former is a prime example of a “boil the ocean”. While this is idealistically this is a desirable alternative, the footprint of EA would most likely outweigh it’s benefits quickly because the of the staffing requirement and high degree of technical oversight.
As a conclusion, there is a right mix of EA governance for federation. The mix is dependent on how an organizations maturity develops and additionally, to what stage the EA practice has evolved to. Governing a federated architecture group clearly takes a minimum number of participants and an organizational structure to keep communications and governance succinct.
