A Few Comments on IBM’s Dynamic BPM
I read Bruce Silver’s post on IBM’s (Websphere) Dynamic BPM recently, and I thought I’d add a few pennies to the information out there. As Bruce points out, the Websphere Dynamic Process Edition is based on Webify, which it just happens was a startup based here in Austin, Texas. In fact, back in 2004 I talked to Webify and got to know a bit about their product, perhaps before all the spin and hype was as well formulated :) Bruce is generally complementary of the offering, and I don’t blame him – its a different approach to processes. But I think IBM and just about everyone else, kind of misses the point of the Webify software. Because IBM puts the word “dynamic” in front of BPM, the focus tends to be on the idea that the process is discovered in real-time based on the customer, product, segment, etc. At the time I talked to Webify, they were pitching the idea of a configurable business process (configurable versus “dynamic”). At the time IBM was pitching “liquid computing” or somesuch and it just seemed a natural fit with an Austin startup that seemed to have already written the key technology to IBM’s marketing spiel. First of all, Configurable is a better (more accurate) modifier than “Dynamic”. Giving parameters that affect the “product” you receive is more accurately described as configured than dynamic (when you use an online car store, and pick your options, it isn’t “dynamic shopping” but depending on what you order it goes to different factories, gets different options installed, taps into different suppliers, etc.). It isn’t that the process is all that configurable, it is that the specific implementation choices for each element of the process are configurable based on the inputs (based on pre-defined policies, pre-defined business service metadata, etc). That’s valuable, no doubt, but the word dynamic appears to modify process, when it really is a modifier against the implementation of specific parts of the pre-defined process. This kind of configuration is powerful, because it can allow the process to select the lowest-cost provider that can still meet SLAs – and the answer to that question may change as the current lowest-cost-provider runs out of capacity, for example – and that change happens “dynamically” if you will, but it isn’t the process that changed – its the selected provider. Bruce puts it well near the end of the post:
WDPE is really oriented to core processes where the degree of variation makes conventional models unmaintainable. In fact, it works best with canonical industry processes where many of the building blocks of solution content can be prepackaged. Not surprisingly, IBM has done that in its key verticals.
I would sum it up this way: most of the BPM tools on the market today are really quite good at representing an organizations internal processes- processes “inside the four walls”. Webify (now, IBM’s Dynamic BPM Edition) seems particularly well suited to processes that live outside the four walls – connecting multiple customers, intermediaries, vendors, suppliers. Hospital systems, insurance companies, drug companies, and the various medical practices involved in providing care for a patient are a reasonably good example of this, and a strong market for IBM’s offering as a result. At the same time, “internal” and customer-facing processes at Blue Cross Blue Shield companies, other insurance companies, and hospital systems are also really good fits for a more “traditional” BPM solution. The Webify / Dynamic BPM Edition approach seems to work best when you can get a network of participants in a particular industry-vertical process to participate in the process implementation. Once you pass the tipping point you can achieve network effects. If you know that some of the founding members of Webify have a background from the old B2B commerce companies like Extraprise/CommerceOne, its no mystery that they would apply some of the lessons learned to a new market and a new software approach. This is pretty neat stuff – and it takes a player like IBM to really leverage the kind of technology that Webify was building out – the scale, the ability to sign up multiple players in an industry into a composite process, etc. I highly recommend Bruce’s whitepaper(free registration required), it goes into considerable detail and provides better background than my meagre comments can, but perhaps these comments will help illuminate the paper further.