Unit cost estimation is the simplest cost-estimation technique: take a known reliable per-unit cost or per-unit revenue, multiply by the number of units, and you have your estimate. Cost = (units) × (cost per unit).
The “unit” is whatever makes sense for the project: square metres of floor space for a building ($1,500/m²), kilometres of road ($2M/km), kilowatt-hours of electricity ($0.12/kWh), days of labour for a consultant ($1,200/day).
The method works well when:
- The per-unit cost is stable and well-known from many past projects.
- The new project is similar in type to the historical projects the per-unit cost came from.
- Scale doesn’t dramatically change per-unit economics. (Buying one truck for $60k doesn’t imply buying 1,000 trucks costs $60M — fleet discounts kick in.)
It works poorly when:
- The new project is unusual in size, complexity, or specification.
- There’s a step change in scale that pushes you across a power-sizing regime.
- The “unit” hides a lot of variation. (/m² for a warehouse.)
Unit cost estimates are typically order-of-magnitude or budgetary in accuracy — somewhere in the ±10-30% range — and are very fast to produce. They show up in early-phase scoping, in budgetary checks during design, and in rough back-of-the-envelope sanity tests for more detailed estimates.
For more precise model-based estimates, see Parametric cost estimation (regression on multiple cost drivers) and Power-sizing model (economies-of-scale exponent). For the broader hierarchy, see Cost estimate classes.