Every construction estimate comes with an accuracy range, whether anyone writes it down or not. A $40 million number produced from a functional program and a cost per square metre isn't the same thing as a $40 million number priced from finished drawings. Yet both often end up in the same board deck, in the same font, with no range beside them. That's how a feasibility figure quietly becomes a fixed budget, and how a project "goes over" before a single drawing is finished.
Canadian quantity surveyors and public owners use a lettered system, Class D to Class A, to label how much an estimate can be trusted. This guide explains what each class means, how it lines up with the AACE International Class 5 to Class 1 system used in many other sectors, and what an owner has to provide to get each one.
The four Canadian classes
The lettered system traces back to the Canadian Institute of Quantity Surveyors (CIQS). Public Services and Procurement Canada uses a four-level version based on it, running from Class D (least design) to Class A (most). The clearest public description is the Guide to Cost Predictability in Construction, published in 2012 by a joint federal government and industry taskforce through the Canadian Construction Association. In summary:
- Class D is based on the initial functional program and a broad concept, usually priced from a rough cost per square metre. The guide puts its accuracy at roughly plus or minus 20 to 30%, depending on complexity. Its main job is to test whether the budget and the program fit together. It often leads to a change in one or the other.
- Class C is prepared at schematic design, about a third of the way through design, once plans, elevations, sections and the main materials are known. The guide stresses that it estimates fair market value, not the low bid.
- Class B is based on working drawings with more detailed dimensions and preliminary mechanical and electrical design. The guide describes it being upgraded through design development, and it can be produced at 50%, 66% or 95% design.
- Class A is based on 100% complete construction documents. The guide expects it to land within 5 to 10% of the median of competitive bids for new construction, and says the estimator needs at least two weeks after the drawings are finished to prepare it.
Firms publish slightly different ranges for the middle classes. One national project management firm, for example, describes Class C as plus or minus 15 to 20% and Class B as plus or minus 10 to 15%. The exact numbers matter less than the principle: the range narrows only when the design does.
How the letters line up with AACE Class 5 to 1
Outside Canadian building practice, the most common system is the five-class matrix from AACE International, also published as ASTM E2516. Class 5 is the least defined and Class 1 the most. Each class is tied to a level of project definition, from 0 to 2% for Class 5 up to 70 to 100% for Class 1, and to typical estimating methods: area factoring and parametric models early, detailed take-offs and unit prices late.
Alberta Infrastructure's Project Cost Framework (2026) maps the two systems side by side. Its expected accuracy ranges are:
- Class 5 / Class D (functional program): −30% to +50%
- Class 4 / Class C (schematic design): −20% to +30%
- Class 3 / Class B (design development): −15% to +20%
- Class 2 / Class B (50% to 75% construction drawings): −10% to +15%
- Class 1 / Class A (95% to 100% pre-tender drawings): −5% to +10%
View the data
| Class (AACE · CIQS) | Design stage | Low | High |
|---|---|---|---|
| Class 5 · D | Functional program | -30% | +50% |
| Class 4 · C | Schematic design | -20% | +30% |
| Class 3 · B | Design development | -15% | +20% |
| Class 2 · B | 50–75% drawings | -10% | +15% |
| Class 1 · A | 95–100% drawings | -5% | +10% |
Two things stand out. First, the ranges are lopsided. The high side is always wider than the low side, because early estimates miss scope far more often than they include scope that later disappears. A Class D number is much more likely to be low than high.
Second, the Canadian letters don't divide neatly into AACE's five boxes. Class B covers a long stretch of design, from design development to well into construction documents. When a consultant says "Class B," ask which drawings it was priced from.
Ranges also vary by sector. British Columbia's transportation ministry, for instance, uses much wider ranges for its highway estimates, from −50% to +100% at Class 5. Infrastructure with large unknowns below ground carries more uncertainty than a typical building.
What each class needs from the owner
An estimate can only be as good as the information behind it. The most common reason an estimate is "wrong" is that it was labelled one class higher than its inputs supported. Before asking for an estimate, owners should know what the estimator needs.
View the data
| Class | Design maturity | What the estimator needs | Used to |
|---|---|---|---|
| Class D | Functional program, broad concept | Space program, site, zoning, $/m² data | Test feasibility; fit budget to program |
| Class C | Schematic design, about 33% | Plans, elevations, outline specs, geotech | Preliminary budget; compare options |
| Class B | Design development to 50–95% drawings | Developed drawings, elemental breakdown | Set construction budget; cost control |
| Class A | 100% construction documents | Complete drawings, specs, 2+ weeks to price | Confirm budget before tender |
The guide also points to what owners themselves must supply. For Class D, that's a final functional program: goals, explicit space requirements, site requirements and regulatory constraints such as zoning. For Class A, it's time. If the drawings are issued on Friday and tenders close the following week, there's no real Class A estimate, just an updated Class B.
Ranges should come from risk, not a table
The ranges above are guides. Alberta Infrastructure's technical guidance on estimate classes is explicit that accuracy should come from a project-specific risk analysis rather than being assumed from the class. Ranges for neighbouring classes can overlap. A simple, repetitive building with good historical data might produce a tighter Class D than a complex renovation at Class B.
This is where contingency comes in. The Alberta framework requires contingency to be risk-based and tied to a risk register, with blanket percentages justified rather than assumed. It also separates design and pricing contingency, which should fall to zero by Class 1 or contract award, from post-contract contingency for changes during construction. Our guide on how much construction contingency to carry covers that split in more detail.
Escalation is separate again. A Class A estimate in today's dollars is still short if construction runs two years. The Alberta framework requires the base cost date and escalation period to be stated, with escalation shown as its own line.
Where owners get caught
On programs we've supported, estimate problems rarely come from bad arithmetic. They come from how the number was used.
- The Class D number becomes the budget. An early figure gets announced, approved or put in a lease negotiation, and every later estimate is judged against it. When the Class C comes in 20% higher, it looks like an overrun. It isn't. It's the range doing what it said it would.
- No class is stated. The estimate arrives as a single number with no basis of estimate, no list of drawings, no exclusions and no base date. Nobody can tell what it covers.
- Exclusions get lost. The guide recommends that owners review inclusions and exclusions with the cost consultant at Class C. Items left out early, such as site servicing, owner equipment or abatement, rarely shrink later.
- Estimates are never reconciled. Each new estimate is compared only with the last one, not explained line by line. Alberta's framework expects a variance analysis when an estimate falls materially outside its expected range compared with later results.
- Tender results are filed and forgotten. The guide recommends analyzing bids against the estimates to build a lessons-learned record. Few owners do it, so the same blind spots repeat.
Approvals and estimate classes
Public owners often tie funding approvals to estimate classes. The federal Treasury Board uses its own indicative and substantive estimate categories for project approval, separate from the D-to-A labels. Alberta requires at least a Class 5 estimate for a needs assessment and business case. British Columbia's transportation guidelines say a project budget should rest on no less than a Class 3 estimate.
Private owners can borrow the same discipline. A sensible gate structure approves feasibility funding on a Class D, design funding on a Class C, and a construction budget only once a Class B is in hand, with the Class A confirming the budget before tender. If the board wants a firm number earlier, it should also approve the range that goes with it.
What owners should do
- Require every estimate to state its class, the drawings and specifications it was priced from, the base date, and its expected accuracy range.
- Report ranges, not single numbers, to executives and boards, especially before schematic design is complete.
- Match approvals to classes. Don't fix a construction budget on a Class D. Set gates where the estimate class supports the decision being made.
- Give estimators the inputs and the time. Finalize the functional program before Class D, and allow at least two weeks for a Class A on complete documents.
- Reconcile each estimate to the last, and compare the Class A with tender results to learn what your estimates miss.
The bottom line
An estimate without a class is a number without meaning. Class D tells you whether a project is worth pursuing. Class A tells you whether the tender will land. Everything in between narrows the range as the design matures. Owners who label estimates accurately, approve budgets at the right gates and report ranges to their boards stop being surprised by numbers that were never as certain as they looked.
Want an independent review of your project's estimates and budget gates? Book a 30-minute call with Benjamin R. Consulting.
Sources: Joint Federal Government / Industry Cost Predictability Taskforce, Guide to Cost Predictability in Construction (2012), via the Canadian Construction Association; Alberta Infrastructure, Project Cost Framework for Alberta Infrastructure Projects (2026); Alberta Infrastructure, Classes of Construction Cost Estimates, Design & Technology Series #19 (2025); BC Ministry of Transportation and Infrastructure, Project Cost Estimating Guidelines (2020); Colliers Project Leaders, From concept to construction: building certainty into project costs; Cost estimate (Wikipedia). Accuracy ranges are indicative and vary by source and project. This article is general information, not financial or legal advice.