In order to achieve a mass-market for Enterprise SOA, there is no question that SAP must convince a LOT of enterprises to rethink their basic business processes. And in order to get a LOT of enterprises to do this, SAP must keep the game fairly simple. Why must SAP keep the game fairly simple? Well, simply because the law of averages (and the underlying “bell curve” or “Gaussian distribution”) says that on the whole, average people are going to be the ones who are in charge of rethinking their enterprises’ business processes. And therefore, if a company wants to energize a lot of “average” people, it has to give them a vision which is within their capacity to realize. (This, BTW, was Ellison’s true genius – he realized that there were aspects of the relational model which even the most technically illiterate CIO/CEO/CFO could grasp, and could therefore be used as part of a sophisticated marketing ploy based on the idea of making such CIO’s/CEO’s/CFO’s feel that they were actually capable of understanding something important about IT, even though this “something” was in fact quite trivial.) But the fact that we all want SAP to succeed in mass-marketing Enterprise SOA does not mean that we must refuse to see other worthwhile “niche” applications for Enterprise SOA that may involve business processes which are not “simple” to rethink and restructure. What do I mean by a “niche” Enterprise SOA application? Well, suppose you have a processing problem in your business that’s an “institutional problem”. By an “institutional problem”, I mean one that can’t be solved by rethinking existing business processes because these existing business processes are so deeply embedded in a particular socio-political-economic matrix that they will never be “rethought” in our lifetimes. (These are usually the kinds of processes that are so outrageously bad that one can only shake one’s head in amazement and cynically ask, with Cassius and Cicero: “Cui bono?”) When one finds such “institutional” problems that are caused by business processes which one cannot “rethink” due to socio-political-economic factors, does this mean that one should conclude that Enterprise SOA cannot be used to ameliorate such problems? Not at all. One merely has to see how the principles of Enterprise SOA can be applied “after the fact” instead of “before the fact”, in order to identify problems after they have occurred and make sure that appropriate steps are taken to rectify these problems. This view of Enterprise SOA proceeds from the assumption that life is messy and human beings are messy and that sometimes, the best Enterprise SOA can do is to help clean up the messes, not prevent them from happening in the first place. In this thread here: Before BI there was AI: will Enterprise SOA bring it back? I mentioned two problems in military logistics that are “institutional” problems in the sense that they occur due to bad business processes that will never be rethought in our lifetimes. I’m sure that there are many similar problems in many other vertical sectors, and that “niche” Enterprise SOA can be used to ameliorate such problems, even when the underlying business processes cannot be re-invented from the ground up.