Mayfield is CEO of Socialtext and has feet (literally) in both the development and business worlds when it comes to real world applications related to Enterprise Web 2.0. I like how he thinks and writes about enterprise adoption of systems and processes based on Web 2.0 technologies.
In a recent posting titled SAPPHIRE Blogger Corps he described the SAPPHIRE 06 conference sponsored by SAP in Orlando. Links to conference information are here (SAP's page) and here (a conference specific wiki sponsored by Socialtext).
Mayfield's blog entry is a very good read -- I recommend you check it out. Here are a couple of the high points:
- Mayfield attended as a "credentialed journalist." He does point out some entertaining encounters with more "mainstream" journalists.
- SAP may be a "stodgy big German company" but Mayfield and others were impressed about the openness to Enterprise Web 2.0 concepts which one might have thought might be anathema to a monolithic legacy system vendor such as SAP.
- Mayfield felt that getting "enterprise bloggers" together with SAP thinkers worked pretty well, even to the extent of successful use of a wiki for recording comments.
I like his comments referencing the inherent disjointness between monolithic architectures and more modular SOA:
First of all, the longstanding claim against a diverse vendor population is that if proliferated results in complexity unwanted by IT. Second, this contradicts the NetWeaver vision of SoA reducing the need to standardize on a software stack and enabling ISVs to deliver value. Third, it's the ecosystem that wags the dog into needed change. Fourth, why would software be the one area of spending that shouldn't adopt a portfolio strategy. Fifth, when I look at the industry today I see substantial innovation.
Ah, yes, it's the classic IT Governance conundrum pitting "centralized control" versus "decentralized chaos." Corporate IT departments, constantly forced to do more with less, are forced into making application consolidation and retirement arguments that lead to ... surprise surprise ... adoption of monolithic mutli-function legacy applications supplied by a single vendor, who then goes on to wag the dog.
Two differences now instead of when ERP and other monolithic architectures were introduce 30+ years ago are that (1) innovative and powerful functionality is available through the web, thereby reducing infrastructure investment requirements, and (2) business units are much more accomplished now in specifying -- and finding and paying for on their own -- what they need technology to do.
The fact that IT departments in larger companies are stuck with maintaining huge expensive legacy applications gives IT departments much less flexibility for being creative and innovative, especially if the costs for investing in these legacy applications are still on the books and can't be easily written off.
Of course, it's unfair to place all the blame for this situation in the lap of IT, as others have pointed out and as I've already documented in my own Web 2.0 Management Survey. Mayfield makes related points elsewhere in his blog:
For example, a primary property of social software is easy group forming -- but most enterprise systems expressly prevent it. To form a group, you not only need permission from IT, but complex configuration and in many cases even software development. Beyond applications, ever come across an LDAP implementation that supports easy group forming? This runs counter to the way many enterprises actually work today, where ad hoc cross-functional teams drive more than professional services organizations.
Given how important relationship development and collaboration are to so many Web 2.0 applications, I think Mayfield has put his finger on an underlying issue -- architecture flexibility -- that will have a lasting impact on how quickly larger organizations can move into more flexible application environments.
My guess is that large companies like SAP are already well positioned to manage a transition to more collaborative functionality built on top of current architectures. Whether such a transition is real or just skin deep I can't tell, but in the corporate world, managing risk is of paramount importance. If a mainline ERP vendor or outsourcer can offer BOTH traditional functionality control as well as reasonable levels of collaborative knowledge and relationship management functionality, risk averse corporate executives should smile.