One of the nice things about working for a large consulting company is that I have access to our strategic services department. These are the people who help develop strategies for our clients and research industry trends and conditions. I recently spoke with a few of our folks who are concentrating on the business-to-business (B2B) market. This discussion was part of what fueled this month's column.
Last year's hottest trend in B2B, the Net market, has cooled down considerably, for a couple of reasons. First is the general trend in the market for business plans with a concrete path toward profitability. In addition, the experience of the past 18 months has shown that one of the early principles for the development of these exchanges was fundamentally flawed.
Neutrality was seen as a prime consideration in the Net market, as it was felt that companies wouldn't wish to put their inventory (and thus a great deal of business intelligence) in the hands of a rival by running an exchange. But time has shown that the liquidity of the market is more important to its success than the neutrality. Liquidity is a basic measure of the community that a Net market attracts to its site. While neutrality is useful, coalitions of large companies are forming exchanges that become either the prime conduit for their business or one of the leading channels. This in turn draws others in the industry to participate because of the increased ability to do so. As my strategist friends like to say, "Liquidity trumps neutrality." A great example of such an exchange is Enron Online, which is generating incredible volume and was one of the first B2B markets to cross over from its primary industry (energy) into other types of business. These two factors have combined to reduce the number of Net markets and have led to a consolidation in this sector.
The market has also seen several large software providers go from a model in which they sold products to an ASP model in which they supply services and host software. ePit is a good example of such a plan. One effect these companies have had is to basically commoditize the trading engine market, even to the point of entering the end business themselves.
All this is driving us toward the next step in the trend - business process engines. We already have software that's approaching this idea, things like EAI and middleware, designed to allow loosely coupled systems to interact. But so far these systems have been largely about communication between applications. The next generation of systems will be about connecting business processes and allowing companies to collaborate within their supply chain to model, adopt, and change their overall processes without the need for intense IT redevelopment.
The vocabularies - languages needed to allow businesses to communicate at this level - are already available or under development. Rosettanet is one; UDDI is another. Most are about defining business processes in a specific context, so processes can be modeled and acted upon.
What's not available but will be in the next stage in the evolution of the business world are systems that will interpret these languages in a way that will allow companies to model their processes in a powerful, easy-to-use fashion. Companies like BEA are working on products like eCollaborate that are approaching this goal.
There's still a rub here. Years ago I helped develop an optimization system for the paper industry. Problem was it was extremeley complex. So is any business of appreciable size.
When you look at a business process, you can usually break it down into many smaller derived processes. It's one thing to define messaging that will allow applications to talk and interact. No matter how you look at it, application-to-application integration still leaves the development of business processes square in the hands of programmers - because they write the applications. It's quite another task to move the responsibility for that definition into the hands of business owners. But that's the ultimate value-add in this space - allowing a business to define how it wants to go to market based on business conditions, not software APIs.
It's still a ways away. The software has to be developed, and more importantly, the business community will have to get used to it. Once that's happened, though, we'll be adding incredible value to the industry. Now that's process improvement.
Sean Rhody is editor-in-chief of Java Developer's Journal. He is also a respected industry expert and
a consultant with a leading Internet service company.