Cost Estimate Classification System (General)

What the classification system is

The single driver behind everything is maturity of project definition — how much of the design and scope is actually known. An estimate made when 2% of the design exists can't be as accurate as one made at 70%, no matter how skilled the estimator. The classification system makes that honest relationship explicit and standard.

Why five classes — and what defines each

Each class corresponds to a band of project definition, and from that flows everything else: how accurate the estimate can be, what it's fit to be used for, and how it's built. Four characteristics distinguish the classes:

1. Level of project definition — the primary driver, expressed as % of complete design/scope definition. Everything else follows from this.
2. End usage — what the estimate is for: concept screening, feasibility, budget authorization, control, or bid/tender.
3. Methodology — how it's built: stochastic/parametric (early) shifting to deterministic/detailed (late).
4. Expected accuracy range — the realistic spread of the final cost around the estimate — wide early, tight late.

The five classes, explored

Select a class to see its maturity, accuracy, purpose, and method. Watch the accuracy tighten as definition grows:

Using the classification well

The system's value is in matching the estimate to the decision. Three habits separate professionals from amateurs:

1. Always state the class. An estimate without its class is incomplete — like a measurement without units. "$10M, Class 4" tells the reader how much to trust it; "$10M" alone invites false confidence.

2. Match the class to the decision. Don't authorize full funding on a Class 5 screening estimate, and don't waste months on a Class 1 estimate when a Class 4 feasibility number is all the gate needs.

3. Don't fake maturity. You can't produce Class 1 accuracy from Class 5 information. If the design isn't defined, the estimate can't be precise — and pretending otherwise sets the project up to blow its budget.

Ten things to remember

1. The classification system grades estimates into five classes by project definition.
2. Level of project definition is the primary driver — everything else follows.
3. Class 5 is roughest, Class 1 is most definitive — the numbers count down to certainty.
4. Four characteristics define a class: definition, end usage, methodology, accuracy range.
5. Method shifts parametric → deterministic as definition grows.
6. Accuracy ranges are honest spreads, not errors — wide early, tight late.
7. Always state the class — an estimate without it is like a measurement without units.
8. Match the class to the decision — don't fund on a screening estimate.
9. You can't fake maturity — Class 5 information can't yield Class 1 accuracy.
10. Industry companions tune the bands — process, building, mining, and more set their own.