The first step in developing a parametric model is to establish its scope. This includes defining the end use of the model, the physical characteristics of the model, the cost basis of the
model, and the critical components and cost drivers.


The end use of the model is typically to prepare conceptual estimates for a process plant or system. The type of process to be covered by the model, the type of costs to be estimated by the model (TIC, TFC, etc.), the intended accuracy range of the model, etc. should all be determined as part of the end-use definition.


The model should be based on actual costs from complete projects, and reflect your organization’s engineering practices and technology. The model should generate current
year costs or have the ability to escalate to current year costs. The model should be based on key design parameters that can be defined with reasonable accuracy early in the project
scope development, and provide the capability for the estimator to easily adjust the derived costs for specific complexity or other factors affecting a particular project.

Data collection and development for a parametric estimating model requires a significant effort. Both cost and scope information must be identified and collected. The level at which the cost data is collected will affect the level at which the model can generate costs, and can always be summarized later if an aggregate level of cost. The scope information should include all proposed design parameters or key cost drivers for the model, as well as any other information that may affect costs.

After the data has been collected, the next step in the process of developing a parametric model is to normalize the data before the data analysis stage. Normalizing the data refers to making adjustments to the data to account for the differences between the actual basis of the data for each project, and a desired standard basis of data to be used for the parametric
model. Typically, data normalization implies making adjustments for escalation, location, site conditions, system specifications, and cost scope.

Data analysis is the next step in the development of a parametric model. There are many diverse methods and techniques that can be employed in data analysis. Typically, data analysis consists or performing regression analysis of costs versus selected design parameters to determine the key drivers for the model. Most spreadsheet applications now provide regression analysis and simulation functions that are reasonably
simple to use. The more advanced statistical and regression programs have goal-seeking capabilities, which can also make the process easier.

Lastly, the resulting cost model and parametric estimating application must be documented thoroughly. A user manual
should be prepared showing the steps involved in preparing an estimate using the cost model, and describing clearly the
required inputs to the cost model. The data used to create the model should be documented, including a discussion on how the data was adjusted or normalized for use in the data analysis stage. It is usually desirable to make available the actual regression data sets and the resulting regression equations and test results.


All assumptions and allowances designed into the cost model should be documented, as should any exclusions.
The range of applicable input values, and the limitations of the model’s algorithms should also be explained.