Towards the end of a project to outsource a large section of a company’s infrastructure I learned that a previous IT Director was very proud of her decision a few years beforehand to insource much of what we were putting back out. Looking into the reasons for both of these transitions you could conclude that there is a right time to have stable operations run at a low fixed cost, and there is a right time to increase the cost, and the risk of cost escalation, in order to allow rapid change by removing barriers and pooling knowledge tightly within a highly skilled team.
Well if both are valid at times, could they not both be valid at the same time? Is there not a case for “rightsourcing”?
The strategic and more-or-less stable systems can be managed in the background by a supplier, only needing major attention for periodic refreshes that replace outdated or capacity-bound technologies but reproduce like-for-like services. The tactical solutions that allow the business to move flexibly can be run by specialised in-house teams who are closer to developing business units and can permit the use of early technology adoption and rapid implementation. This way the costs and abilities are matched to the need for change, so the tactical vs strategic split sounds like a practical theory.
But in reality, how would you co-exist these two opposing models? There is obvious potential for moving tactical point solutions into the strategic mainstream over time, so service agreements would have to be supple enough to support this. There is a risk that tactical solutions might need transforming before they can be transitioned out into the supplier’s support model, so the in-house teams would need to adopt stricter standards early enough avoid cost barriers when they want to outsource a stabilised system.
The greatest challenge is likely to be choosing how to carve up technology layers that could be common to both camps. For instance, you probably want to have only one network, so you would consider this strategic. However tactical solutions might need certain components of the network to flex and adapt as they tried to make funky new solutions work across the company. But could you realistically say that the backbones and switches are managed by a vendor yet the firewalls are managed in-house? How would the supplier manage to SLAs when there are dependencies on in-house teams for core components of the layer they are trying to manage.
This is not to say that its impossible to carve up responsibilities between the two camps – simply that it will take a lot more skill and forethought to structure relationships that allow harmonious co-existence rather than leaving process gaps that can fill with excuses or blame.