[To gain context, read Ismael Ghalimi's BPM Inferences for 2006. This is my reply.]

Ismael,

Nice post... too bad you're completely misreading the market!! :-)

The value-add standard of BPM isn't BPEL, it's BPMN and, ultimately, executable BPMN (which may be in the form of a merger of BPMN and UML, then executable UML). Ultimately, rationalizing the business language of BPM will be done via a graphical metaphor, not more code.

Therefore, the technology stack vendors (as you imply) will own BPEL as that form of commoditization occurs, but the business notion of "business process management" (lower case) will be won by the application vendors (SAP, Oracle and Microsoft) and one or two "independent" process-oriented vendors (horizontally building process-foundations and, subsequently, offering multiple vertical frameworks as they grow). These application vendors understand business values, as opposed to IT values, and will work to expand bpm via the visual metaphor of the process, not another language. Native execution drives zero business value, as long as interoperability exists. For example, it doesn't matter whether I write a service in Java or .NET, frankly, as long as the service is discoverable and interoperable. In the same way, it doesn't matter what code is used to execute a process - as long as that execution is driven by a graphical model I can easily change and exchange in standard ways. (I'll leave to a different day the discussion of how the selection of a given language adds value... it typically adds value not because it is inherently better, but simply because you use it consistently and there is a talent pool from which to recruit, so that training costs are lower... this is why BPEL is years away from proving itself as a long-term cost-effective language... there simply hasn't been enough time to know whether a talent pool will really develop around it. In my opinion, more likely is that a different, more generalized BPEL-like language will come along, and that will be the next step. By the way, the .NET and Java camps will, in fact, drive a wedge in this "standard" and embrace extensions... same as it ever was...).

Let me make a different prediction for 2006: More adoption will be driven by business and IT alignment than by the language choice of the IT department. And, that business/IT alignment will be highest in companies using a consistent visual metaphor at design-time, during simulation & optimization and during process execution. Companies that focus on the visual model will advance faster than companies who focus on the language.

By the way, I agree that most of the so-called BPM vendors will go away, doing what they have always done: building custom applications for a vertical they know best. I do not agree that the best of the pure-plays will go this way. The best of the pure-plays that do not remain independent will be purchased, primarily, by application vendors who want to have great BPM technology inside their existing application suite, and who feel that proprietary execution methods offer them strategic advantage (this may or may not be true in specific cases, but it's why there will be acquisitions and not simple embedding of open-source alternatives). It's possible that a stack vendor like a BEA or Sun would buy a pure-play, although those companies would be better-served catering to their constituency with a native BPEL engine. Hard-core advocates of the technology stack will prefer BPEL. Hard-core advocates of the business stack will prefer BPMN. The latter will win, in the end, just as it has been with ERP (SAP, for example, just passed the $10 billion revenue mark... whereas BEA's total revenues are about 1/10th that. Being kind to IBM WebSphere and Oracle's App Server, that makes the app server market about $3 billion total, 1/3 the size of only SAP... I am not counting MS Windows Server, which is a significant web platform, too, but you get the point...).

And one more thing... it could be that the "independent BPM" winner(s) are not the companies they are today. Acquisitions will be made so that acquiring companies can enter this huge market, and then become the independent winner. But there will be at least one behemoth who is not identified with today's application or technology stacks..... might as well be Lombardi, right?

Cheers and see you at the next OMG event,
Phil

Technorati Tags: BPEL, BPM, BPMThinkTank, Intalio, BPMN, UML