In my last post, I started the discussion of managing IT costs by defining why you need to get a good handle on costs (beyond the obvious reasons).

It's important to understand the total service costs because understanding those costs can drive operational improvements. So let's start by looking at IT-service-based costing.

What makes up a service cost?

Let's revisit the equation from my previous post:

Total Service Cost = Initial Service Project Costs + Ongoing Maintenance and Support Costs

Initial Service Project Costs: First, you need to actually build the service out and typically there's a project associated with that, so your initial service project costs need to be factored into the costs of the services (these are typically amortized over the life of the service investment).

Capital expenses: These relate to all of the hardware, software and applications that contribute to this service. (These are typically amortized over the life of asset). These costs can be directly linked to the service.

Cost allocation can get complicated when assets are shared across services. In the case of dedicated servers, the total cost can be assigned to a service. Where resources are shared, however, we need to move to a service-allocation model based on resource consumption. Also indirect (overhead) costs like physical computer facilities need to be proportionally allocated across services.

Operating expenses: There are some fixed expenses (like labor costs, maintenance, service enhancements, etc.) that can be attributed to a particular service. But again, if support staff manages across services, how do we assign these costs to a particular service? We need to move to a service-allocation model based on consumption.  This should include variable service support costs related to how well a service is behaving because that directly relates to the support costs. For example, if a service is continually failing because of infrastructure problems, software bugs, etc., someone in IT has to spend time (labor costs) to fix those issues. So we want to start attributing some of the support costs based on incidents (in the service desk) against the service, and use the time taken to fix those incidents by the labor costs to the estimate these variable costs. We should also capture all maintenance and service enhancements as Requests for Change and cost each change.

We can make broad sweeping macro decisions that can affect our CAPEX (e.g., server consolidation) and OPEX (e.g., reduce employee head count), which will have an impact across all services. But we can't direct operationally improvements or evaluate service value. Enter IT-service-based costing:

IT service-based costing (aka, activity-based costing for IT)

Activity-based costing is a costing model that in which you assign costs for resources of a given activity out to the products and services requiring that activity, based on the actual consumption by the products and services. By using this methodology, you can assign more of your indirect (overhead) costs to direct costs.

This allows you to establish the true costs of your products and services so that you can better assess their value (and determine whether you should continue investing in them or not). You can also use the resulting activity cost data to determine where to focus your operational improvement efforts.  In IT the "products and services" are IT services and similarly we can direct operational improvements in IT by assigning costs to the IT services provided. 

Service consumption: This directly impacts your variable costs. It permeates down to the individual assets that underpin the service.  So here is a key insight: If we use an activity-based costing model, we want to assign the costs based on consumption and use models that assign asset costs that are shared across services proportionally by service. To do this, we need an accurate map of the assets that are configured to each service. Here is where a CMDB can be utilized to do more than it was originally intended to do.  We can use service configuration information in the CMDB to help aggregate the individual service costs to each service. And if an asset is shared across services, we can divide the costs proportionally based on consumption of the asset by the service.

So now let's revisit our equation again:

Total Service Costs = Amortized Service Project Costs

                             + Sum of all service consumed resources costs

                             + Sum of all service variable service costs

                             + Sum of all service maintenance and enhancements

                             + Sum of all overhead costs allocated to the service

So we need to adjust our thinking a little. There are some things we can do at an IT macroeconomic level to improve our cost situation across all services, and there are other things we can do at a microeconomic and operational level to improve things for specific services.

Come back soon and we'll investigate how to make both macro- and microeconomic improvements to your cost structure in my next blog post.