We can assist in providing answers in the following areas:
- organisational performance measurement;
- costing/benefit review of policy options;
- portfolio simulation/optimisation;
- data discovery/acquisition;
- costing policy implications;
- project/organisational cost modelling;
- regulatory cost models; and
- quantitative risk analysis.
Cost Modelling
There are two common approaches to modelling they are "bottom-up" and "top-down". Both are used to provide cost estimates. A bottom-up approach takes details and aggregates them to produce a cost estimate. A top-down approach takes a budget and then removes details to produce a cost estimate. Ideally both models produce the same results. However, typically both modelling approaches will miss some details. The top-down approach which is removing costs will tend to not remove enough cost and end up overestimating aggregate costs. Conversely the bottom-up approach misses out some elements and will end up underestimating aggregate costs. This is often seen in project costs. In a bottom-up model every discovery about the costs tends to increase them.
cost model purpose
Cost models exist for many reasons including but are not necessarily limited to: budget forecasting (a time horizon of perhaps 1-2-3 years); economic whole of life modelling (time horizon may be 40 or more years), and regulatory impacts. Each of these model types is different and requires different emphasis.
cost models
A robust cost modelling approach consists of:
- Determining the platform that the model will be built on. Perhaps the most common platform is Microsoft Excel. The choice of platform should go beyond what is convenient or what is at hand to consider the need for traceability, change control, speed of calculation, and updating the input data.
- Identify the cost drivers: - these are the parameters that when changed affect the cost. Consideration should be given to whether a parameter will have a significant impact on the cost. Nobody would ever suggest that models be cluttered with insignificant cost drivers. In a complex model often the significance of a driver is not obvious until the model is built.
- Establish how the parameter affects the cost - does it interact with other parameters? What is the nature of that interaction? An example of this might be the unit cost of fuel, the rate of consumption of this fuel, and the hours of operation of the platform. All three factor interact to produce a cost.
- Construct the model.
- Validate the model, preferably the validation and an external review will occur with someone who has not been directly involved in the production of the model.