43R-08Intermediate12 min read

Parametric Estimating — Example Models (Process)

A beginner's walkthrough of a parametric contingency model applied to a process-industry project — scoring the systemic risk factors, running them through the model, and reading out a contingency. Built on AACE International RP 43R-08.

Why process industries?

Parametric contingency models were pioneered on process-industry projects — refineries, chemical plants, LNG, pharmaceuticals — and that's where they're most mature. These projects share enough structure (defined phases, repeatable equipment, decades of cost history) that the relationship between a project's characteristics and its overrun can be reliably modelled. RP 43R-08 provides example models for exactly this setting.

Scoring the systemic factors

A typical process-industry parametric model rates the project on several factors, each on a defined scale (here, a simple 1 = best to 5 = worst). The model's fitted coefficients then weight each factor's contribution to contingency:

Systemic factorWhat it measuresScore (1–5)
Scope definitionCompleteness of scope & engineering at estimate3
Process technologyHow proven vs novel the process is2
Project complexityNumber of systems, interfaces, constraints4
Project size / locationScale and site difficulty3
Team & data qualityExperience and quality of cost basis2

Worked example: a process plant

Take a $50M process plant with the factor scores above. An illustrative model assigns each factor a contingency contribution; summing them gives the total. (Real models use validated regression coefficients — these are round numbers for teaching.)

$50M process plant — parametric contingency build-up

Worked example
Base allowance (all projects)4%
Scope definition (score 3)+5%
Process technology (score 2)+2%
Project complexity (score 4)+5%
Project size / location (score 3)+1%
Team & data quality (score 2)+1%
Predicted contingency18% → $9.0M

The model predicts 18% contingency, or $9.0M on the $50M base. Notice what drove it: scope definition and complexity together contributed 10 of the 18 points — the systemic factors doing exactly what Lesson 12 said they would.

Lessons from the example

Use a model validated on projects like yours, score the factors honestly and ideally independently, and present both the total and the factor contributions so decision-makers see what's driving the number. Treat the result as the systemic-risk contingency, and reconcile it against expected-value or Monte Carlo estimates of the discrete risks — when the two diverge a lot, that gap is itself worth investigating.

Nine things to remember

  1. Parametric models are most mature on process-industry projects.
  2. You score the project on systemic factors, not individual risks.
  3. Key factors: scope definition, technology, complexity, size, team/data.
  4. Worked result: 18% contingency → $9.0M on a $50M plant.
  5. Scope definition and complexity dominated the contribution.
  6. The scoring is where the judgment — and bias — lives.
  7. Read the factor contributions — they point at how to reduce contingency.
  8. Use a model validated on similar projects and score independently.
  9. Reconcile against line-item methods — big gaps are worth investigating.

Glossary

Base allowance
Contingency every project carries.
Coefficient
Fitted weight turning a score into contingency.
Contribution
Each factor's share of total contingency.
Factor score
The rating given to each systemic factor.
Process technology
How proven vs novel the process is.
Reconciliation
Comparing methods' results for consistency.
Scope definition
Completeness of scope and engineering.
Systemic risk factor
A scored project characteristic driving contingency.

Check your understanding

1Parametric models are most mature for which projects?
2In the worked example, the model predicts what contingency on a $50M plant?
3In parametric scoring, the judgment (and bias) mainly lives in: