Rightsizing Service Cost Models, Part IUse these theories from ITSM Watch guest columnist Frank Bucalo of CA to design effective and efficient cost models.
This article focuses on cost models for ITIL Financial Management. It reviews ITIL cost modeling theory, and makes recommendations to apply the principles above using real-world use cases.
Cost modeling in ITIL consists of identifying cost elements and their costsdiscrete cost factors that comprise your servicesand fairly allocating those costs to the services they support. Cost types include: hardware, software, personnel, accommodation, external (non-discrete costs from external service providers), and transfer (non-discrete costs from internal departments).
Costs are further categorized as:
The Use Case
The use case for this article is a Web-based, distributed application in a virtualized environment. This complex use case covers almost every cost type including multiple cost element layersthe real elements and the virtual elements. (Warning: Never start your ITIL implementation with such a complex service! Start simple, learn, and grow.)
The most common approach is to measure everything everywhere. In ITIL terms, this means making every cost element either direct or absorbed via a usage metric. As outlined in the initial article, this approach is bottom up and ignores considering the cost-benefit ratio of measurement.
In our use case, this effort would be huge. We would have to determine the usage of the distributed hardware and software including servers, network, storage and floor space as well as the time allocated to each task by a system administrator.
Once we measure everything at the physical level, we would need to do the same at the virtual level. Then we would have to associate the usage with the appropriate users and apportion costs weighted by the usage metrics. Clearly, the problem is that we are at the wrong level of granularity. This approach will most likely bog down and ultimately fail.
A better approach is to minimize the required measurements. Start with the simplest method you can find and drill down only until an acceptable level of fairness is achieved at an acceptable cost.
In our use case, we might start with measuring the number of services (e.g., two) sharing the virtual environment, treating all costs as a single unabsorbed cost (e.g., $100). Each service would then be assumed to cost $50 ($100/2). This clearly isnt the final solution and is patently unfair, but by approaching the problem from this direction, you can come to a fair and equitable service cost apportionment algorithm in the most efficient manner.
For example, one could use the number of service users (Service A=75 users, Service B=25 users) as a cost apportionment weighting factor. This metric is fairly easy to capture and could increase the fairness of the apportionment. Now we would cost service A at $75 and service B at $25.
Next, we might want to use the number of application page hits (e.g., Service A=30, Service B=70) as the weighting factor. This metric starts to capture the level of system usage, potentially adding more fairness, still at a reasonable cost. Service A would be costed at $30, and Service B would be costed at $70.
This article illustrated how taking a simple top down approach to service costing makes developing the cost modeling feasible. My next article will elaborate and illustrate some specific principles for defining more complex and potentially fairer cost apportionment algorithms.
Frank Bucalo is a senior architect at CA. Frank has more than 20 years of experience implementing business applications for the Wall Street community. Over the last five years, Frank has a track record of successfully delivering ITIL implementations from business analysis, through intelligent design, and technical implementation.