Two estimators using the same unit rate can still price the same trench differently. The fix is pairing every rate with a fixed, written measurement convention, not trusting that everyone means the same thing by it.
Key takeaways
STEP 01 OF 10
CCDC 4 prices work "for a pre-determined, fixed amount for each specified unit of work performed. The total price is determined by multiplying the unit price by the actual, measured quantity of work performed." Two variables sit inside that sentence, not one: the rate, and the measurement method. "The same unit rate" means different things in two estimators' hands unless both variables are standardised together.
This matters most exactly where it looks least necessary — a rate everyone thinks is obvious, like a common trench excavation or a standard drywall installation, is precisely where an unstated measurement assumption tends to hide, because nobody has ever had a reason to write it down.
STEP 02 OF 10
Is a linear metre of trench measured neat-line or including overcut? Is a square metre of drywall measured net of openings or gross? A rate book that states the dollar figure without stating the measurement convention beside it is not actually standard — it just looks standard until two estimators compare notes on the same scope.
A short note beside each rate is enough: the item, the unit, the measurement rule, and the date it was last confirmed. It does not need to be elaborate to close the gap that causes most unit-price disputes between estimators.
STEP 03 OF 10
Tie this to the supplier pricing feed’s reconciliation step: when a reconciled invoice shows a rate has genuinely moved, version the change in the rate book with a date and a reason, rather than silently overwriting the old number. An estimator working from a six-month-old printout should still be able to see what changed and when.
Versioning also protects a firm defending a completed job: if a client or a subtrade disputes a rate applied eighteen months earlier, the versioned record shows exactly what rate was in force on the date the work was priced, not just what the rate happens to be today.
STEP 04 OF 10
The same accountable-owner discipline used throughout this hub — see the AI adoption plan for the same principle applied to a different decision — applies here. One named person approves any exception to the standard rate book, and the exception is logged with the job it applied to.
An exception that is never logged has a way of quietly becoming the new normal for the next estimator who copies an old job as a starting point, without anyone deciding that it should.
STEP 05 OF 10
CCDC 16 names "methods for adjusting the contract price" as one of the core subjects its change-order guidance covers. When a unit-price item's scope changes mid-contract, route the rate adjustment through the same documented change-order discipline described in building an estimating system that scales, rather than negotiating a one-off number that never makes it back into the rate book.
A rate negotiated once, informally, for a single change order tends to resurface as a precedent on the next job whether or not it should — routing it through the documented process is what stops a one-off concession from quietly becoming a new standard rate.
STEP 06 OF 10
StatCan’s monthly releases on investment in building construction are one honest trigger point: pick your own internal threshold for how much movement in the relevant month's figures should prompt a rate review, and re-check the book on that schedule rather than waiting for a bid to come in wrong first.
Combine this with the supplier feed's own reconciliation cadence rather than running two unrelated review schedules — a rate book review that happens the same week as the pricing feed reconciliation catches drift in both places at once.
STEP 07 OF 10
Pull a handful of recent estimates and check every unit-price line against the current, versioned rate book. This catches an estimator quietly working from an old printout, or a rate that drifted through informal exceptions that were never logged, before either becomes a pattern.
Pick the sample at random rather than reviewing only the estimates that already seem to have gone well — the point of the audit is to find the drift nobody has noticed yet, not to confirm what already looks fine.
STEP 08 OF 10
Never delete a superseded rate outright — supersede it, with the date and reason kept in the record. Federal record-keeping rules already require six years of retained business records tied to your tax filings, and a five-year-old estimate should still be explainable from the rate book that produced it, not from memory.
A rate book that only ever shows the current figures cannot answer the question "what did we assume when we priced this job," which is exactly the question that comes up when an old project is reviewed, disputed, or used as the basis for a new estimate.
STEP 09 OF 10
A firm operating in more than one province cannot use one holdback assumption across the rate book. Ontario, British Columbia and Alberta all retain 10%; Manitoba's Builders' Liens Act s. 24(1) retains 7.5%. A rate book that inherits its cash-flow assumptions from whichever province the firm started in will misprice every Manitoba job by a real, computable amount — not a rounding difference.
Mark the Manitoba exception directly on the rate record, the same way step one already asks estimators to write CCDC's own definition next to the rate. An assumption that only lives in one senior estimator's head is exactly what step four's exception process exists to eliminate.
STEP 10 OF 10
StatCan's Building Construction Price Indexes, second quarter 2026 reported non-residential costs up 2.7% in Québec City and just 0.2% in Vancouver over the same period — both cities, not provinces, and both from the same release used for step six's re-basing. A rate book that escalates "Québec rates" and "BC rates" as two provincial blocks is coarser than the data actually reporting.
Where the release names a city your firm bids in, use that figure. Where it doesn't, fall back to the province or the national number and record which fallback was used, so a future audit under step seven can see the reasoning, not just the resulting rate.
Assuming a shared spreadsheet is the same thing as a shared standard. A spreadsheet everyone can open is not the same as a rate book everyone actually uses the same way. Without the measurement convention from step two, two estimators can read the same cell and price the same scope differently.
Copying an old estimate as a shortcut without checking the rate version. Reusing a prior estimate as a starting point is efficient, but only if the rates it carries are checked against the current versioned book first — otherwise a superseded rate quietly rides into a new bid.
Letting exceptions accumulate without ever reviewing them as a set. A single logged exception is manageable. A dozen unreviewed exceptions across different jobs is often an early sign that the standard rate itself needs updating, not that every job is unusual.
Treating the audit in step seven as a formality once a year. An audit that happens once and finds nothing tends to stop happening. Keep it on the calendar twice a year regardless of what the last audit found.
Standardising the rate but not the units it is quoted in. A rate book that mixes metric and imperial units, or per-square-metre and per-square-foot pricing, across different trades reintroduces the same inconsistency the standardisation effort was meant to remove. Pick one convention per trade and hold every estimator to it.
Building the rate book once and treating it as finished. A rate book is a living record, not a one-time project. The versioning, exception-logging, and re-baselining steps above only work if the book keeps being actively maintained well after the initial effort of building it has faded from memory and the novelty has worn off.
Blending the escalation index and the holdback rate into one 'provincial adjustment.' They move independently, per the comparison above, and a rate book that conflates them cannot show which one actually changed when a number needs correcting.
Step nine names the outlier. Here is what happens if it isn't caught.
Scenario A. A $2,000,000 Manitoba subcontract, priced from a rate book that (incorrectly) carries the 10% holdback assumption from the firm's Ontario projects, shows $200,000 retained against progress payments.
Scenario B. The same subcontract, priced correctly against Manitoba's actual 7.5% rate, shows $150,000 retained — $50,000 less. A subcontractor pricing against the wrong figure either walks away from a bid that was never as cash-tight as the book suggested, or the firm quietly overstates its own retained-capital position by $50,000 on a single subcontract.
This is the same shape of error as the province-merging trap the rest of this hub warns about repeatedly — one province's number, carried somewhere it doesn't apply, priced in real dollars rather than a technicality.
A unit rate book scaling across provinces is really carrying three separate external numbers under one heading of "provincial adjustment." Keep them visibly separate.
Blending these three into one "Manitoba adjustment" factor on the rate sheet hides which one actually changed when a rate looks wrong later. Keep them on three separate lines.
The 7.5% figure in the Manitoba Act applies to the value of work and materials retained under the contract structure generally — treat it as the default for the province rather than assuming it applies only to labour, and confirm against the specific contract if a supplier's terms differ.
Yes — fold it into the same twice-yearly sample audit rather than creating a separate review cycle just for provincial exceptions.
Apply each province's own rate to the portion of work actually performed in that province, per the contract's own allocation — this is not a case for picking whichever single rate is simpler to apply.
A 30-minute call is enough to tell you whether it is worth building.
