Phase 2: Cost allocation is optional and a business decision. Cost allocation - if done - should be as simple, fair, predictable and controllable as possible.
At first sight, allocating the costs of shared services looks easy. You simply divide the service cost by the number of users. But both parameters are open to question. On service cost, users will come and count the number of servers, the number of tapes, and so on. And the number of users rarely reflects usage. What do you base it on? Number of PCs? Wall jacks? There will always be disputes about the numbers, and usually you need to apply a little more sophistication.
Seven methods are now commonplace.
High-level allocation (HLA) is based on a simple business parameter. HLA allocates costs by a simple business parameter, such as business unit revenue or number of employees. Defining the scale of a business unit, assumes a level of service used.
For example, take three business units with annual sales of $100m, $200m and $300m respectively. The cost of $6m for operating a data centre would be allocated as $1m, $2m and $3m.
HLA works best for shared services (like infrastructure) where the lack of a clear relationship between usage and cost doesn't matter.
Low-level allocation (LLA) is based on an IT-related parameter. LLA allocates the cost of IT services using an IT-related parameter, such as number of PCs or network IDs. But like HLA, LLA favours some users.
For example, treasury departments in banks and exploration departments in resource companies form the bulk of the enterprise and are technology-intensive. Other business units in these organisations may have a similar number of PCs on desks, but nowhere near the usage. Yet they would still have to pay a similar amount for their technology, as the resource-intensive departments.
LLA works best for services like desktop support and IT architectural activities, where a suitable parameter can be readily agreed and everyone has similar levels of use.
Direct costs (DC) are allocated to a single owner. With DC, all the costs of a service, regardless of usage, are allocated to the service's single owner. This makes it unsuitable for shared services.
DC is often used for clearly attributable services, such as application development and dedicated projects.
Measured resource usage (MRU) measures IT resources. MRU allocates the costs of a shared IT service by accurately measuring the IT resources. Each business unit pays for what it uses. The most common example is voice communications because the provider measures usage.
MRU is best suited to services with a high-variable cost component, such as data storage and telecommunications.
Tiered flat rate (TFR) charges using levels. TFR allocates costs by service tiers (service bands), with each level charged at a fixed (flat) rate. Take the example of providing technical support. A one-hour response, same-day response or next-day response could each be charged at a different rate.
TFR is best suited to stable environments where service levels are predefinable, like help desks, application maintenance and data centres.
Negotiated flat rate (NFR) is based on annual usage. NFR, a refinement of TFR, is based on a business unit's annual service usage. IS and the business unit then negotiate a flat monthly fee. For example, some business units negotiate for help desk and technical support.
NRF is most suited to stable environments, such as help desk, technical support and infrastructure services. Don't use NFR when change and uncertainty are the norm.
Market-based pricing (MBP) aims to make a profit. Unlike the other methods, which aim to recover costs, MBP aims to make a profit. It assumes that the IS organisation operates an open market inside the enterprise and sells services to business units at market prices. It's an "IS Pty Ltd". Business units should be given the choice of sourcing services from IS or from an external service provider.
Note: Gartner research in this area shows that TFR, MRU and LLA will have widespread use across a range of IT services, in all industries, by 2005.
Once costs are allocated, the third and final stage in chargeback is to recover them.
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.