Document control is not a filing habit, it is the system that decides whether the right drawing revision, the right submittal status, and the right invoice are still findable six years from now.
Key takeaways
STEP 01 OF 10
Bluebeam's own interactive stamp templates for submittal review use three core workflow statuses — Approved, Revise, and Rejected — alongside fields for reviewer, date, submittal number and spec section. Adopt that exact vocabulary as the firm standard rather than letting each project invent its own status labels; a submittal log with six different words for "needs changes" across six projects is not a system, it is six habits.
Whatever software marks up the submittals, the underlying status set and its required fields should be identical project to project. That is what lets a document controller move between projects without relearning the log each time.
STEP 02 OF 10
A revision letter ("Rev C") tells you the sequence but not the date it superseded the prior version — and a dispute months later often turns on exactly when a revision was issued and received, not just which one is current. Log both the issue date and the date each recipient acknowledged it.
Keep superseded revisions accessible, not deleted. A document control system that only shows the current revision cannot answer the question "what did the drawing say when this work was actually built."
STEP 03 OF 10
Federal law, not a construction standard, sets the real retention requirement: under ITA s.230(4)(b), records must be kept "until the expiration of six years from the end of the last taxation year" they relate to — and under s.230(4.1), electronic records must stay in an electronically readable format for that whole period. This clock runs from the tax year, not from substantial completion or the warranty period, which is why a document control calendar built only around project milestones misses it.
Where no return was filed for a year, s.230(5) means the retention clock for that year's records never starts — those documents effectively have to be kept indefinitely. Flag any project with a filing gap for permanent retention rather than the standard six years.
STEP 04 OF 10
Under ITR 238(2), any person or partnership whose business income is derived primarily from construction activities and that pays subcontractors must file a T5018 information return, on either a calendar-year or fiscal-year basis (once chosen, that choice cannot be changed without ministerial authorization). The return is due within six months after the end of the reporting period — a hard filing deadline that depends entirely on subcontractor payment records being complete and organized well before it arrives.
Build the T5018 deadline into the same document-control calendar as submittal and drawing tracking, not a separate accounting-only system. The two are only separate in name — the underlying records are the same subcontractor payment documents this whole system exists to control.
STEP 05 OF 10
CCDC's landing pages describe what its standard forms cover — CCDC 2, for instance, addresses "the role and authority of the consultant, procedures for changes in the work, work by other contractors, insurance requirements, prerequisites for Ready-for-Takeover, dispute resolution procedures" — but the actual clause text is sold, not published. A document control system should reference the public description of what a clause covers, never a specific clause number pulled from memory or a secondary source. See CCDC 2's own description.
This matters because CCDC 2's own description states the contract "assumes monthly progress payments" and references provincial payment legislation where applicable — language a document control system's payment-application tracking should mirror, not contradict.
STEP 06 OF 10
Ontario's O. Reg. 213/91 requires a Notice of Project, a Form 1000 registration, trench notices for excavations over 1.2 metres, and asbestos notices, each with its own filing rule and posting requirement. Treat these the same way as a submittal — logged, dated, and posted at the required location — because a Ministry inspector asking to see a Notice of Project is asking exactly the kind of document-control question this system exists to answer instantly.
See from award to mobilisation in one week for when these notices actually get filed in the project sequence.
STEP 07 OF 10
A document control system with no single accountable owner degrades into whoever last touched the folder structure. Name the role explicitly, even on a small project where it is a part-time responsibility layered onto a PM's existing job — the point is that someone is accountable for the log being current, not that the role requires a full-time hire.
The named backup matters as much as the primary — a document control system that stalls the moment one person is on vacation is not actually controlling anything.
STEP 08 OF 10
A short weekly check — open submittals against their required response date, open RFIs against their due date, this week's drawing revisions logged and distributed — catches a missed item while it is still cheap to fix. The alternative, a scramble at closeout to reconstruct what happened over the previous year, is where document control systems actually fail: not because the documents were never captured, but because nobody looked at the log until it was too late to matter.
See building a quality control programme for how this same weekly discipline extends to inspection and test records specifically.
STEP 09 OF 10
A document control system that requires a site supervisor to drive back to a trailer or an office to log a photo or a daily note is a system that gets used inconsistently. See rolling out digital daily reporting for the same principle applied specifically to daily logs — the mechanism is identical: reduce the friction between the field and the record, or the record stays incomplete.
This does not require expensive software. A phone photo logged with a date and a location tag, filed into the correct project folder within the day, beats an elaborate system nobody uses consistently.
STEP 10 OF 10
At project closeout, archive the submittal log, the drawing revision history, the OHS notices, the RFI log and the T5018-supporting payment records together, indexed to survive the six-year ITA retention clock from step three. See project closeout in thirty days for how this archiving step fits into the broader closeout sequence.
A closeout binder that only contains the final drawings and the warranty certificates is not a complete archive — the six-year clock applies to the underlying records, and those are the ones a document control system was built to keep findable.
Deleting superseded drawing revisions instead of archiving them. A dispute about what was actually built often turns on which revision was current on a specific date — that question is unanswerable once the prior version is gone.
Setting the retention calendar around project completion instead of the tax year. ITA s.230(4)(b) runs the six-year clock from the end of the taxation year the records relate to, which rarely lines up neatly with a project's own end date.
Inventing project-specific submittal status labels instead of a fixed, firm-wide vocabulary. A reviewer moving between projects with six different status vocabularies will eventually misread one, and a misread submittal status is how the wrong revision gets built.
Treating the T5018 deadline as purely an accounting responsibility. The underlying subcontractor payment records are the same ones document control already tracks — splitting ownership just means nobody notices the six-month deadline until it has passed.
Waiting until closeout to reconcile the document log. A weekly ten-minute check catches a missed RFI response cheaply; a project-end scramble to reconstruct a year of records rarely recovers everything.
The T5018 filing deadline and the ITA retention period are both date-driven, not judgment calls — here is what they actually produce for a firm with a calendar fiscal year.
A GC's fiscal year ends 31 December 2025. Its T5018 return covering that year's subcontractor payments is due within six months of the reporting period's end — 30 June 2026 at the latest. Under ITA s.230(4)(b), the records supporting that return — the subcontractor invoices, the payment ledger, the T5018 itself — must be kept until six years from the end of the 2025 tax year: 31 December 2031. A document filed on the last day of a project that closed out in early 2026 could still need to be produced nearly six years later, in a compliance review that has nothing to do with the project itself.
Neither date depends on the project’s own timeline — both are set by the tax year, which is exactly why a document control calendar built only around project milestones will miss them.
Document control has an unusual shape compared to the rest of this cluster: its hardest deadlines are federal, not provincial — but the notices it has to log are provincial, and they are not the same notices everywhere.
A document controller working across provinces needs both calendars — the federal one for retention and filing, and the provincial one for which notices apply and when they are due.
A disciplined shared-folder structure with a fixed naming and status convention can work for a small firm — what matters is the consistency of the vocabulary and the weekly reconciliation habit, not the specific tool.
Responsibility for the filing can sit with a bookkeeper or accountant, but the underlying records still need to be captured and organized by the document control system in real time — a bookkeeper cannot file an accurate return from an incomplete log.
That is a compliance gap with the CRA, independent of any construction-specific consequence — it is a strong reason to build redundant, off-site or cloud backup into the archiving step rather than relying on a single physical or local copy.
Keep the core three — Approved, Revise, Rejected — fixed across every project. Additional project-specific fields can be layered on top, but changing the core vocabulary itself is what causes cross-project confusion.
A 30-minute call is enough to tell you whether it is worth building.
