Construction Budget Overruns Start Before the Site: The Estimate Accuracy Problem

A construction budget does not break in the field. It breaks in the weeks before the field, when the estimate is prepared, the scope is written and the assumptions are locked. When the overrun appears on site, everyone points at the crew or the contractor. But the overrun was only reported on site. It was decided earlier.

This article focuses on one of the least understood causes of overruns: the accuracy of the estimate itself. It explains how estimate accuracy depends on project definition, why generic and optimistic estimates fail, and how to build an estimate that matches the decision it supports.

Every estimate has an accuracy range

Professional cost engineering recognises that estimate accuracy depends on how well the project is defined. AACE International’s Recommended Practice 18R-97 classifies estimates from Class 5 to Class 1:

• Class 5 (concept screening): very little definition; typical accuracy ranges from about -20% to -50% on the low side and +30% to +100% on the high side.
• Class 4 (feasibility): limited definition; roughly -15% to -30% and +20% to +50%.
• Class 3 (budget authorisation): design partly developed; roughly -10% to -20% and +10% to +30%.
• Class 2 (control or bid): substantial definition; roughly -5% to -15% and +5% to +20%.
• Class 1 (check estimate or bid): nearly complete definition; roughly -3% to -10% and +3% to +15%.

The lesson is simple: a budget approved on a concept-level estimate is not a fixed number. It is a range, and the range is wide. Problems start when an early estimate is treated as if it had Class 1 accuracy.

Three ways estimates go wrong

1. Generic, not project-specific

Estimates based on cost per square foot from past projects or published guides ignore what makes this project different: site conditions, existing building constraints, access, specifications and complexity. They are useful for screening, not for budget approval.

2. Optimistic, not precise

When the estimate is prepared to make the project look viable or to win a client, allowances shrink and contingency disappears. The project starts with a number that was never realistic.

3. Disconnected from the market

Pricing that ignores current conditions (trade availability, material lead times, tariffs and escalation) becomes outdated before construction starts. In Alberta, Altus Group’s 2026 Canadian Cost Guide reports overall cost increases of 4% or more in Calgary and Edmonton, and labour shortages continue to reduce the number of bids.

Well-scoped projects absorb surprises

Every project encounters surprises. The difference is how much room the budget has to absorb them. A well-defined project with an accurate estimate and a risk-based contingency absorbs a hidden pipe or a delayed delivery. A vague project with an optimistic estimate cannot, and every change order compounds the previous one.

How to build an estimate that matches the decision

Match the class to the decision

Use concept estimates to screen options, feasibility estimates to choose a direction and budget-level estimates only when design is sufficiently developed. Do not approve a fixed budget on a screening estimate.

Develop the scope before pricing

A written scope, measured surveys and drawings developed to the appropriate level are the inputs that make an estimate accurate.

Price with current market input

Involve contractors or cost consultants with current pricing. Ask key trades for budget quotes on major packages and long-lead items.

Include escalation and contingency explicitly

Show escalation to the procurement date and contingency linked to identified risks as separate lines, not hidden inside unit rates.

State the accuracy range

Every estimate presented to an owner should state its class and expected range. This simple habit changes how budgets are approved and managed.

Review the estimate independently

An estimate checked by someone not invested in the project’s approval catches optimism and omissions.

When the decisions come too late

Many budgets die because key decisions arrive after they could still change the outcome: finishes chosen during construction, equipment selected after rough-in, structural details resolved on site. The cost of a decision rises sharply as the project progresses. Making decisions early is the cheapest form of cost control.

An example of estimate drift

Consider a commercial fit-out approved at a concept-level estimate of $1.2 million. During design development, the client adds a second meeting room, upgrades lighting and requests acoustic treatment. Meanwhile, mechanical equipment quotes come in higher because of tariff-related pricing on metal components. None of these changes is unusual, yet by tender the realistic cost is well above the approved figure. Because the original number was never presented as a range, the owner experiences the difference as an overrun rather than as expected refinement. Presenting the first estimate as a Class 4 or Class 5 figure, with its range, would have set expectations correctly and triggered earlier decisions on scope.

How to present estimates to owners

A simple change in presentation prevents many disputes. Instead of a single number, show the estimate as a range with its class, list the main assumptions and exclusions, and identify the three or four items most likely to move the number. Owners then understand that the figure will be refined as design progresses, and decisions about scope or finishes can be made while they still affect cost.

The bottom line

Execution cannot save a bad estimate. Pre-construction can. For anyone running a project, the question is not only whether the team can execute, but which scope, which assumptions and which reading of the market the budget is built on. By the time the crew arrives, the math is already locked.

If you want a second read on an estimate before it is locked, see my pre-construction and project management services, read about why budgets fail before site, or contact me.

Frequently asked questions

How accurate is a construction cost estimate?

It depends on project definition. Under AACE International’s classification, early concept estimates may vary from about -50% to +100%, while estimates based on nearly complete design are typically within about -10% to +15%.

What is an estimate class?

A category describing the maturity of project definition and the expected accuracy range of an estimate, from Class 5 (concept) to Class 1 (detailed).

Why do construction estimates fail?

Because they are generic rather than project-specific, optimistic rather than precise, or disconnected from current market pricing and escalation.

Sources

• AACE International, Recommended Practice 18R-97: Cost Estimate Classification System
• Altus Group: 2026 Canadian Cost Guide insights

• 110 views
• 1 Comment