A quality control programme that only produces a passing average is not actually controlling quality — the COR audit structure proves the point on its own numbers.
Key takeaways
STEP 01 OF 10
COR's national audit instrument works on a per-element basis: a minimum of 65% in each element, and an overall audit score of 80%. Structure the firm's own internal QC programme the same way — separate scored categories for inspection frequency, non-conformance closeout, training records, and so on — rather than one blended quality metric that can hide a genuinely weak area behind a strong overall number.
IHSA is named as the Authority having Jurisdiction to grant COR in Ontario, and the programme itself began in Alberta more than 20 years ago before being adopted nationally through the Canadian Federation of Construction Safety Associations — useful context for why the audit structure is consistent across provinces even where the administering body is not.
STEP 02 OF 10
Name, in advance, what gets inspected, at what frequency, and what document records the inspection — a checklist with a signature and date, not a verbal walk-through. This is the same discipline the document control system in a document control system for a GC applies to submittals, extended to the physical work itself.
Where a technical standard governs the inspection — CSA standards, for instance — note that CSA's own standards documents sit behind a paywalled store, not published free. Where a firm cannot cite the specific clause, cite the National Building Code 2020's relevant chapter through the NRC instead, or describe the requirement generically and direct the reader to confirm the current provincial code locally.
STEP 03 OF 10
CCDC 2's public description names "prerequisites for Ready-for-Takeover" as one of the contract's core mechanisms — a free, citable anchor for what a quality programme should be checking before a phase or trade is signed off as done, even without access to the full clause text. Build the QC checklist's sign-off criteria around this concept rather than an informal "looks finished" standard.
This same Ready-for-Takeover concept is what feeds directly into project closeout — see project closeout in thirty days for how a QC programme's own sign-off records shorten that final sequence.
STEP 04 OF 10
A non-conformance report that is logged but never confirmed closed is a QC record that looks complete and is not. Require a closeout signature on every non-conformance, with the corrective action described, before the associated work is covered up or the phase is signed off.
This is the same principle as COR's own audit structure: identifying a problem is not the same as scoring well on it. A programme that logs issues but rarely closes them will show that gap the moment anyone audits the closure rate specifically.
STEP 05 OF 10
COR audits treat training as its own scored element, not an afterthought to technical inspection. Track certifications, refreshers and toolbox-talk attendance the same way inspection checklists are tracked — dated, signed, and filed where an auditor or a client can find them without a scramble.
For Ontario's 23 compulsory trades, this overlaps directly with the licence verification already built into a subcontract package template set — a Certificate of Qualification check at award is the front end of the same training-record discipline a QC programme tracks through the life of the project.
STEP 06 OF 10
A programme scoring 90% on most elements and 62% on one is failing, under COR's own structure, regardless of what the overall average shows — because 62% sits below the 65% per-element floor. Build the internal scoring the same way: report every element's own number, not just the blended average, so a weak area cannot hide behind strong ones elsewhere.
This is a genuine trap in any quality system that reports a single headline number. A programme that only tracks the average will not notice it has a failing element until an external audit finds it first.
STEP 07 OF 10
Inspection results, non-conformance status and toolbox-talk attendance should feed into the same daily reporting stream covered in rolling out digital daily reporting, not a separate QC-only log nobody else sees. A QC record that only the quality manager can find is a QC record that is not informing the site's day-to-day decisions.
This also means the same weekly reconciliation habit from document control applies here: a QC log reviewed weekly catches a pattern of repeated non-conformances in one trade while there is still time to address it with that trade directly.
STEP 08 OF 10
No Canadian regulator or industry body publishes site-theft loss rates, trade-by-trade waste factors, or equipment utilisation benchmarks that a QC programme could cite as a target. Where a programme wants a numeric target for these, build it from the firm's own historical data, and label it explicitly as an internal benchmark — never present a circulating industry figure with no traceable Canadian source as if it were a verified standard.
This restraint is itself a quality-control discipline: a QC programme built on invented benchmarks fails the same way an unverified submittal status does — it looks authoritative and is not.
STEP 09 OF 10
If the firm intends to pursue COR certification, the daily and weekly discipline in the steps above — per-element scoring, closed non-conformances, current training records — is what an audit actually reviews. IHSA requires pre-payment for the audit review, and no fee figure is published on its own site, so confirm current cost directly with IHSA rather than quoting an old number.
A firm that only starts building these records the month before a scheduled audit is building a paper trail from memory. A firm running the programme daily is simply exporting what already exists.
STEP 10 OF 10
A QC element that consistently scores low because the schedule never allows time for proper inspection is a scheduling problem wearing a quality label. See standing up a scheduling function for how inspection and hold points should be built into the schedule itself, not squeezed in around it.
This last step is the one that turns a QC programme from a compliance exercise into an actual management tool: it asks whether the schedule is setting quality up to succeed or fail before the first inspection even happens.
Reporting a single blended quality score instead of per-element results. COR's own structure fails a programme with one element under 65% regardless of the overall average — an internal system that hides that behind a headline number will miss the same failure.
Logging non-conformances without a required closeout signature. An issue that is identified but never confirmed closed is a record that looks complete and is not.
Citing a CSA standard's clause number from memory or a secondary source. CSA standards sit behind a paywalled store — without the text in hand, cite the National Building Code through the NRC or describe the requirement generically instead.
Using an invented or vendor-sourced benchmark for waste, theft or utilisation rates. No Canadian body publishes these figures — a programme that needs a target should build it from its own historical data and label it as internal, not industry-standard.
Running QC as a parallel system the rest of the site never sees. A QC log only the quality manager checks is not informing daily decisions on site, which defeats the purpose of building it at all.
COR's two-part threshold — a 65% floor per element and an 80% overall score — produces a result that a single-number quality metric would miss entirely.
An audit scores eight elements: 90, 85, 88, 92, 62, 95, 89 and 84 — an overall average of 85.6%, comfortably above the 80% requirement. Under a single-number system, this programme passes easily. Under COR's actual rule, it fails, because one element — scored 62% — sits below the 65% floor every element must individually clear. The other seven elements being strong does not offset the eighth being weak.
The lesson generalizes past COR specifically: any quality programme that only tracks a blended average will not notice a genuinely weak area until it is measured on its own, which is exactly why the internal scoring in step six of this guide reports every element separately.
COR is a single national audit instrument, but who grants it, and what feeds into meeting the underlying safety-committee obligations behind it, differs by province.
A firm running QC programmes across provinces should check the local committee threshold before assuming Ontario's 20-worker rule applies — Saskatchewan in particular will require the underlying safety-committee structure well before other jurisdictions do.
No — it is a voluntary, nationally trademarked audit standard, though some owners and general contractors require it as a bidding prerequisite on specific projects.
IHSA's page names the certification body and audit structure but does not publish a fixed recertification interval in the material reviewed here — confirm the current cycle directly with the administering body for the province.
Yes — the per-element scoring structure, the non-conformance closeout discipline and the training-record tracking are useful on their own, independent of whether the firm ever pursues formal certification.
IHSA's own page states that cost information is available by contacting IHSA directly and does not publish a figure — do not state a price without confirming it current for the specific audit type and firm size.
A 30-minute call is enough to tell you whether it is worth building.
