This topic may seem dated. While the tipping point for cloud services has come where many enterprises have completed or are planning to move the majority of their platforms to the cloud, higher education is looking at the balance of service delivery between central and academic and auxiliary units.
This topic can be viewed in similar terms whether traditional cloud or hybrid / internal cloud. Re-evaluation of services, self-hosted, or purchased is something that should constantly be done. Large institutions of higher education often have multiple IT departments. Growing out of a tradition of open access to information and having unfettered access to the tools necessary for research and teaching, technology tools were implemented before a mature IT management structure could be there to guide the organization as a whole.
Things evolve and change happens. An institution will want to re-evaluate its structure which is wise from time to time. The key is how. Any change imposed on units with their own autonomy can be a touchy process. There are universal values, universal strategies and then there are local values and strategies. There is no reason rational players cannot agree in principle on changes mean value to an org. To help the process, here I show a categorization of services and a process that helps all parties come to an agreement on what to look at when it comes to consolidation.
First, the categorization. I place services in three tiers. The idea is to stratify services by their ability to be commoditized. The first tier (1) are commodity services. Email, file & print, network backbone, virtual stacks are all low hanging fruit examples where customization and a managed tie to the operation it serves is not necessary. These services should be offloaded and most organizations would be more than pleased to do so. Skipping tier 2 for now, tier 3 services are those that are not easily offloaded or serve not efficiency advantage. Project Management and business process management are examples. Web design and areas that intersect with marketing and design are others. Many IT departments perform hybrid functions and those usually are unique that offer no efficiency or alignment advantage to consolidate and in fact, will likely do some disruptive harm if so. Dispensing with the low hanging fruit at both ends, we turn to tier 2. These services are the ones where there will either be disagreement or some sort of judgment call about the efficacy of a move. For that, I provide a decision making rubric and a process that allows you to make a decision on a service by service basis. The central IT group or central administration do the scoring from their perspective and the unit IT group/administration do the same scoring from their perspective. The scoring items on a scale of 1 to 10 are: 1 – how intimate is the structure and relationship of the service to the organization it serves, 2 – how much does the service contribute to the goals of the org, 3 – how much value does it contribute to the business, 4 – what is the balance of risk vs reward for moving (this is a spectrum score where risk scores high and reward scores low). These factors are multiplied by each other and a score of 1 to 1000 results. The two scores are added and averaged and you have a final answer. You can pre-determine what score suffices a decision recommendation, say 500 and below is a consolidate recommendation.
After the scoring is complete, a decision workflow asks some questions. First, can the central IT org perform the function better than the local one and if so, does the central org feel it is of value to the institution to make the consolidation move. If the answer is yes to both then there are three more questions to answer. To what extent should consolidation take place? What will the approach be, will local unit expertise be required? Finally, what staff resources both local and centrally will be required for a smooth transition.
There you have it. Now, look at examples in the real world where consolidations have happened and have been as smooth as the above methodology promises. Do you know of one? I’m interested in knowing about your observations.
Michael Dawisha 