Article details
Every EPC (Engineering, Procurement, and Construction) project starts with a number nobody fully trusts yet. Early on, that number might be off by 50% in either direction — and that's not a failure, it's simply how cost estimating is supposed to work at that stage. The real skill isn't producing one perfect number upfront; it's knowing how confident an estimate should be at each stage, and refining it deliberately as the project matures.
Here's how that actually plays out, following a single project's estimate as it evolves from a rough guess to a number the whole project gets built on.
Stage One: The Rough Estimate That Decides Whether to Proceed
At the earliest stage — before detailed drawings exist, before equipment is selected, sometimes before the scope is even fully defined — a project needs a number just to decide whether it's worth pursuing further. This is where a parametric or analogous estimate comes in: using statistical models tied to a few known variables (building size, system type, climate zone) or leaning on costs from comparable past projects, to produce a rough-order-of-magnitude figure fast.
Industry classification systems — most notably AACE International's estimate classification system — formalize this staging. At this earliest stage (commonly called a Class 5 estimate), accuracy ranges as wide as -50% to +100% are normal and expected. This isn't sloppiness; it's an honest reflection of how little is actually known yet. Anyone treating a Class 5 number as a firm budget is misunderstanding what that number is for.
Stage Two: Scope Solidifies, and the Estimate Tightens
As design progresses — engineering disciplines finalize their scope, equipment gets selected, layouts take shape — the estimate needs to mature alongside it. This is typically where estimating shifts from purely analogous comparisons toward incorporating more genuine project-specific detail, blending historical cost data with an increasingly defined scope.
This stage is also where a hard truth about estimating shows up clearly: a vague or incomplete scope is the single most common cause of a bad estimate, more than any error in the estimating methodology itself. A perfectly executed cost estimating process applied to an incomplete scope — a missing piece of routing, an unaccounted-for process step, an assumed-but-unconfirmed equipment selection — will still produce a number that's wrong, because the estimate can only be as complete as the scope it's built on.
Stage Three: The Detailed, Bottom-Up Estimate
By the point design is finalized — equipment schedules locked, drawings complete, installation sequencing understood — the estimate shifts to a bottom-up approach: quantifying every actual cost component rather than extrapolating from comparisons. This is the most labor-intensive estimating method, but also by far the most accurate, typically landing within roughly -15% to +20% of final cost when done well.
At this stage, a complete estimate accounts for four distinct cost categories, and missing any one of them is a common source of budget overruns:
Direct costs — labor, equipment, materials, and subcontractor costs tied directly to construction and installation
Indirect costs — project management, site facilities, administrative overhead — the costs that don't show up as a line item of physical work but are genuinely required to execute the project
Contingency costs — a deliberate allowance for the unknowns that inevitably surface: unexpected site conditions, design changes, coordination issues between trades
Escalation costs — an allowance for material and labor cost increases over the project's duration, which matters more the longer a project's timeline runs
Summing these four categories establishes the cost baseline — the reference point everything else during execution gets measured against.
Why the Estimate Doesn't Stop Once Construction Starts?
A mistake worth naming directly: treating the estimate as finished once construction begins. In practice, the estimate should keep being revisited at defined project milestones — a stage-gate approach — updated against actual costs incurred, any scope changes, and current market conditions for materials and labor.
This ongoing discipline is what actually catches a small deviation before it becomes a major overrun. A project team that only checks costs at the very end has given up the chance to correct course while correction was still cheap.
Choosing the Right Estimating Method for the Stage
Three methods cover most of what's needed across a project's lifecycle, and choosing the wrong one for the stage is itself a common estimating mistake:
Parametric estimating — statistical, variable-driven models; useful early, when detailed data doesn't exist yet
Analogous estimating — leaning on historical costs from genuinely comparable past projects; fast, but only as good as how comparable those past projects really are
Bottom-up estimating — building the total from fully quantified, itemized components; the most accurate, but only viable once scope is genuinely well-defined
Using a bottom-up approach too early wastes effort on a scope that hasn't stabilized yet. Relying on a parametric estimate too late into a project, when detailed information is actually available, leaves real accuracy on the table unnecessarily.
Why This Matters Beyond the Cost Engineering Team?
Cost estimating is often treated as a discipline handled entirely by dedicated cost engineers or project controls teams — and on larger projects, it often is. But any engineer contributing to project scope benefits from understanding why an estimate is at a certain stage, and what it can and can't be trusted to predict. Knowing that a late design change carries real estimating consequences, or that an unclear piece of scope will come back as a cost surprise, is engineering judgment as much as it is cost engineering.
Further reading: EPC Cost Estimating: Keys to Accurate and Successful Projects, Nomitech.
