In a related discussion to my previous post Nick Malik and Joe discuss approaches to delivering SOA within the organisation. I am wholly onside with Nick’s notion that the key advantage of SOA is the ability to federate decisions about implementation out to the individual service providers and that the role of enterprise planners is to facilitate connections across these providers in a light touch manner. As a result I believe that the identification of the capabilities required by an organisation – along with their responsibilities and metrics – belongs with enterprise planners but that anything from this point on is federated out to service providers (whether internal or external). Essentially I feel that the best way to manage service orientation in a sustainable way is to concentrate on metricisation and monetisation and leave the rest to the ‘market’ (whether internal or external); essentially let self interest flourish within the bounds set by the organisational context as long as it delivers cost-effective services but punish it by outsourcing where it doesn’t.
As I understand Nick’s description of his view of middle out I believe that it is similar to the combination of top down and bottom up that I’m describing here. I think that this is on the money as neither top down nor bottom up are sufficient on their own; you need top down to deliver the systematic view of the enterprise required to support adaptability and effective sourcing but at the same time need to empower people at the edges (i.e. the bottom) to deliver on their commitments rapidly and in a way that meets their needs. As a result I believe that the best approach is a combination of governance and anarchy – govern the commitments assigned to the atomic parts of the business but delegate all further implementation concerns to the owners of these atomic units (and this obviously cascades downwards given the fractal nature of services). In this context there may be many duplicated services implemented but this is OK; where there needs to be duplication to support independent operation or because individual services are not economically viable for implementation by a third party there can be, whilst for services that can economically be shared joint ventures or enterprising third parties will emerge to profitably fill these spaces in a way that is sustainable (i.e. the services can be profitably sold and thus sustained rather than just being an overhead and a source of drag on each provider – which is where the link back to my previous post pops in). This approach contrasts with some of the more naive approaches to service-orientation I have seen where people obsess about forcing business units into the use of shared services that are not actually economically viable and which – at the same time – prevent them from adapting independently in response to market changes. Essentially services become an onerous tax imposed by IT rather than a real enabler for change and value creation.
Decide what you need and then depend on the endless ingenuity of your colleagues, partners, customers and suppliers to deliver a response without trying to control how they deliver. You’ll unleash a whole lot more innovation, much more rapidly – and that will be to everyone’s benefit.