Methods & Tools

Efficient billing of application costs

by Jeremy Smith

Billing for applications and their operational costs has always been a concern for IT organisations and their customers. In most cases, the models used are either too simple or too complex, which can lead to dissatisfaction. We show how an efficient compromise can be achieved.


There is a wide range of models and practices to bring transparency into application operating costs and their billing. The goal is often to find the right compromise in the pricing model in terms of fairness, transparency and the effort required (including data collection and allocation). But how do you allocate a team of 200 IT professionals as fairly as possible across your applications without doubling the size of your controlling team?

In my experience, many IT organisations shy away from application billing. They are either unfamiliar with the processes involved or find them too complex. And the remaining companies? One part of the IT organisations charge for the software landscape on an application-specific basis with a high level of detail, while the others charge a flat rate for the costs incurred. Both methods have their drawbacks:

  • The detailed accounting marathon is extremely time-consuming for business units and IT, is not much fun and often leads to high communication costs.
  • Flat-rate invoicing is not transparent and makes it difficult to control and optimise IT and costs, for example, because it does not show the levers with which the business unit can reduce its application costs.



Categorising Business Applications into Effort Classes

Over the past few years, we have had good experience with customers who have adopted our billing model based on effort classes. This is - admittedly - a deliberate compromise. The model defines several effort classes into which the company's applications are then sorted. It does not matter whether a tool is home-grown or off-the-shelf. The categorisation is based on the characteristics of certain cost drivers in the operation of applications. These include

  • Number of APIs
  • Number of compute instances (server, container)
  • Number of users
  • Number and type of changes implemented per year
  • Number and type of incidents handled per year
  • Service quality (e.g. in terms of service time and availability)


Balancing efficiency, transparency and fairness

Our projects have shown that around 95 per cent of applications can be grouped into just four to six effort classes. This approach provides organisations with a simple model that is very close to the actual effort required, and also enables control and optimisation based on the underlying measurable complexity parameters.

In addition, the cost of implementing the charging model can be minimised by focusing on parameters that have a significant impact on the cost and complexity of applications. This applies, for example, to the number of interfaces with other systems and to the number of changes that are implemented. In this respect, effort classes represent a perfect compromise between efficiency and fairness of allocation - a good starting point for companies that do not yet allocate their application operations internally, but also an approach for optimising their own existing process.


ERP systems are left out

Highly complex applications in the organisation can pose a challenge when categorising them. Centralised and globally rolled-out ERP systems are usually calculated individually (often based on defined complexity drivers), as their parameter values often exceed the value ranges of the four to six defined standard classes, resulting in distortions in the effort/complexity metric.

And categorisation into effort classes is only the first step: at least as important is the question of how the costs of applications are subsequently allocated to different business units or national companies. This is where it becomes individual again - but that's another story.



There is no useful universal approach to calculating the operating costs of applications. As a result, many companies resign themselves or opt for one extreme: the very detailed calculation of expenses at the level of individual applications or a flat-rate billing of the entire application landscape. Our pragmatic approach offers a balanced solution between simplicity and complexity and can help companies to make the operating costs of their applications transparent, control them and optimise them in a systematic manner.

Jeremy Smith

Jeremy Smith

Jeremy is responsible for UK, Benelux & Northern Europe and has been in the IT benchmarking arena for over 25 years. He previously received bench­marking exercises as an end user and delivered benchmarking exercises as a project manager.