A common complaint about legacy IT is that it is organized into silos. This doesn't just affect legacy systems, but it may also affect new projects.
In his classic book on the evolution of buildings (How Buildings Learn), Stewart Brand showed how the townscape often lasted longer than the buildings themselves. In an already built-up area, new buildings typically occupied the same footprint as the old building it replaced. The same is often true of new IT systems - the functionality may extend into previously unoccupied areas, but it is more difficult to make radical changes to the boundaries between systems. I have a lot about this in my 2001 book on the Component-Based Business.
But if the sponsorship and funding for new IT projects emerges from a host organization, and if the host organization is itself organized into functional silos, then it is hardly surprising when this structure is reflected in IT. This is an example of Conway's Law.
So we have long argued that new modes of IT sponsorship and funding need to be developed if an organization wishes to get maximum benefit from SOA. But what about the effects on the host organization itself?
There is now a strong pressure within the business world for silo-busting. This is nothing to do with IT, and everything to do with making business operations more effective and joined-up. Ranjay Gulati, a professor at the Harvard Business School, published an article on Silo-Busting last year (Silo Busting: How to Execute on the Promise of Customer Focus, Harvard Business Review, May 2007). He argues that "companies claim to offer customer solutions, but most aren’t set up to deliver them without specific changes in organizational structure, incentives, and relationships". I'm looking forward to reading his forthcoming book on the same subject.
A quick search for Silo Busting finds the idea applied to several areas of management.
- Healthcare ("health care organizations suffer from internal functional barriers that get in the way of achieving the level of success that they want and need"): Leta Beam (April 2006 pdf), R. Wade Schuette (September 2006)
- Marketing: Roy Young (April 2008)
- Retail: Eric Olafson (September 2007)
- Supply Chain: Robert Bowman (September 2008)
So this is a common theme within the business literature, and SOA offers strong support, especially for multi-channel operations.
Gulati identifies four success factors for silo-busting
- Coordination (Structure and Process - this is the first area where SOA and BPM can help)
- Cooperation (Sharing - this is where Web 2.0 or Enterprise 2.0 can help)
- Capability Development (this is perhaps where improvisational comedy comes in)
- Connection with Customers (there is a secondary role for SOA in providing joined-up customer-facing services)
Conway's Law doesn't work in reverse. You can't just build joined-up systems and hope that the desired organizational change will follow automatically. There is always going to be a larger organizational change programme, with a significant but not exclusive role for SOA (and related activities, especially Service-Oriented Business Modelling). So it is hard to quantify exactly how much business value is created by the SOA piece. But we may be able to get some clues from the organizational change projects people are already doing ...