Over the last three years I have met around ten people who have assured me that they originated service integration. At first these claims irritated me, then they amused me but now I realise that they all support the need for and validity of service integration as a model. It has evolved naturally from a changed environment to meet new needs.
What’s striking about the claims is that the various independently designed models are remarkably similar. As, sadly, are their shortcomings.
These shortcomings occur in two main areas.
First, there is little or no recognition of the transformational aspects of service integration. They have been designed bottom-up by operational people who needed a way to manage an existing or impending multi-sourced environment. The focus is on getting control and making sense of the supply; the demand side is largely ignored.
The demand side is where service integration can really make a difference. The end-to-end visibility of IT and business series that it enables allow a closer alignment of the IT organisation with its business customers and their plans than ever. This is increasingly important with the advent of digital transformation and bi-modal IT.
The second area relates to managing the performance of service integration itself. Simply considering the effectiveness or otherwise of end-to-end services is inappropriate; it may or may not be a result of service integration. I have recently seen a suggestion that the percentage of “non-compliance” would be a good measure. Since this would also require knowing the number of compliances, it is an impractical measure. Worse, even if it was measurable it would not tell you anything about the impact or meaning of those non-compliances.
A more meaningful approach is to consider the impact of service integration on business outcomes. If we accept that the primary responsibility of service integration is to assure the performance of end-to-end services then we can move towards measures such as “How many cars did we not make today because of IT service failure?” or “How many customers did not complete an online purchase during a period of poor service performance?”
These kind of measures have to be based on comprehensive data, powerful analytics and human interpretation.
Will this represent the next stage in service integration’s evolution?