A price list updated once a quarter is already behind the market it is supposed to describe. Here is how to build a feed that keeps up, and keeps a defensible, dated record of every figure on it.
Key takeaways
STEP 01 OF 10
StatCan’s release on investment in building construction for May 2026 recorded the national total edging down 0.3% to $23.4 billion for the month, up 5.9% year over year, with provincial swings inside that single month running from a $52.5-million increase in British Columbia's multi-unit segment to an $80.1-million decrease in Alberta's. A feed refreshed quarterly is already several of these swings behind by the time anyone notices.
The same release showed the residential and non-residential components moving in opposite directions inside the same month — residential down, non-residential up slightly — which is exactly the kind of divergence a single blended price list would hide from an estimator pricing one or the other.
STEP 02 OF 10
Monthly, at minimum, matching how often the underlying national series itself is re-released. This is a cadence discipline your firm sets for itself, not a rule any regulator mandates — but matching your refresh rate to the rate the market itself is measured at is the simplest defensible standard to set.
A trade with genuinely volatile input costs may warrant a faster cadence than the general rule; a trade with stable, slow-moving costs may not need monthly attention at all. Set the cadence per category of supplier, not as one blanket rule for every line on the feed.
STEP 03 OF 10
Link the specific supplier quote or catalogue date to each line on the feed, the same discipline used for every fact in this hub: a figure, its date, and its source, every time. A price on your feed with no date attached is not a price you can defend at the negotiating table, or explain to your own estimator six months later when the number has moved.
This also protects the firm the other way: when a supplier disputes a rate the estimate was built on, a dated, attributable record settles the argument in minutes instead of turning into a dispute about who remembers what.
STEP 04 OF 10
This is the same accountable-owner discipline used in the AI adoption plan: one person, named, with the authority to override, and a log of every override with a reason. A feed anyone can quietly edit is not a feed, it is a shared spreadsheet with an unearned reputation for accuracy.
The log entry should record the reason, not just the change — "supplier verbally quoted a lower rate for volume" is a very different kind of override from "correcting a data- entry error," and an estimator relying on the feed later benefits from knowing which one it was.
STEP 05 OF 10
The Excise Tax Act requires that "every person that carries on a business or is engaged in a commercial activity in Canada... shall keep all records" needed to determine tax liabilities and obligations, retained "until the expiration of six years after the end of the year to which they relate," and kept "in English or in French." Supplier pricing records that feed your GST filings and job costing fall under this obligation, not just under your own convenience — six years is the floor, not a suggestion.
A feed that automatically archives superseded prices rather than overwriting them satisfies this by design; a feed that only ever shows the current number needs a separate, deliberate archive step built in alongside it.
STEP 06 OF 10
A supplier's list price and the price your firm actually pays after terms are two different numbers, and a feed that conflates them will misprice every estimate that relies on it. Keep both fields, and make the unit-rate consequences explicit — see standardising unit rates across estimators.
Volume discounts, payment-term discounts, and project-specific negotiated rates all belong in the "your price" field, clearly separated from the published list — otherwise an estimate can end up pricing at list on a trade where the firm has never actually paid list.
STEP 07 OF 10
Not annually. A quarterly reconciliation catches drift between what the feed says and what suppliers are actually invoicing before an estimate gets built on a stale number, rather than discovering the gap at year-end when the affected bids are long since submitted.
A reconciliation that finds no drift at all for several quarters running is itself useful information — it may mean the feed's refresh cadence from step two can safely slow down for that supplier or category.
STEP 08 OF 10
The whole point of the feed is that the estimating system pulls from it directly — the same logic behind a cost catalog feature, where standardized pricing is stored once and reused during takeoff — rather than an estimator working from memory or a personal spreadsheet that nobody else can audit.
If an estimator ever needs to override the feed for a specific job, route that through the same accountable-owner and logging discipline from step four, rather than a silent local edit that never makes it back into the shared record.
STEP 09 OF 10
Statistics Canada's Industrial product and raw materials price indexes, June 2026 release recorded the Raw Materials Price Index down 6.9% month over month but still up 20.7% year over year, and named softwood lumber specifically: prices up 6.1% in the month, "its largest month-over-month increase since November 2024," the sixth consecutive monthly gain for wood products generally. That volatility is exactly why step two's cadence has to be matched to the actual product, not set once for the whole feed.
A lumber-heavy line moving 6% in a single month needs a materially shorter refresh cycle than a fixtures line that hasn't moved in a quarter. Use a release like this one to set the cadence for the categories that actually behave this way, and leave the slower-moving categories on a longer cycle rather than refreshing everything on the same clock.
STEP 10 OF 10
Income Tax Regulation 238(2) requires an information return — the T5018 — from "every person or partnership that pays or credits, in a reporting period, an amount in respect of goods or services rendered on their behalf in the course of construction activities," where construction income makes up the majority of that reporting period's business income. The reporting period is fixed once chosen under s. 238(3) and the return is due "within six months after the end of the reporting period" under s. 238(4).
The quarterly invoice reconciliation from step seven is the natural place to also flag which supplier payments belong in that year's T5018 filing — catching it there, against real invoices, is cheaper than reconstructing it from the pricing feed months later at filing time.
Confusing 'the spreadsheet was updated recently' with 'the prices are current'. A feed can be touched every week and still carry stale numbers if nobody is actually reconciling it against real invoices. Activity is not the same as accuracy.
Letting one category's volatility set the cadence for everything. A fast-moving material cost does not mean every line on the feed needs weekly attention. Matching cadence to category, from step two, avoids both stale slow-moving lines and wasted effort on stable ones.
Storing prices without the measurement unit that goes with them. A price with no attached unit convention is exactly the trap covered in standardising unit rates across estimators — the two problems are closely related and worth solving together.
Building the feed as one person's personal file. A pricing feed that lives on one estimator's laptop is not a system the firm can rely on if that person is away, or leaves. Build it somewhere the whole estimating team can see and audit.
Refreshing the feed but never reconciling it against real invoices. A feed that is dutifully updated from supplier catalogues but never checked against what the firm actually paid can drift quietly for months. The reconciliation step in step seven is what catches that drift; the refresh step alone does not.
Treating every supplier the same way regardless of how much of the cost base they represent. A supplier accounting for a small fraction of total spend does not need the same monitoring intensity as one behind a firm's largest and most volatile cost line. Weight the refresh and reconciliation effort toward the suppliers whose prices actually move the bottom line.
Assuming the feed is only an estimating tool. A well-kept, dated supplier pricing feed is also part of the documentary record behind the firm's own costing and tax filings, which is exactly why the six-year retention discipline from step five belongs in the same system, not in a separate archive nobody checks.
Setting one refresh cadence for the whole feed. A category that moved 6.1% in a single month and a category that hasn't moved all quarter do not belong on the same refresh clock, per step nine.
Step one already argues a static list goes stale. Here is the number behind that argument, not just the claim.
Scenario A. A framing package priced at $40,000 using a list price refreshed at the start of the month absorbs June 2026's reported 6.1% softwood lumber increase in full — $2,440 — if the feed isn't refreshed again before the material is ordered.
Scenario B. The same package, priced from a feed refreshed weekly against the categories flagged as volatile in step nine, catches the increase in the first week it appears rather than the last, giving the estimator the chance to reprice or requote before committing rather than absorbing the full month's movement as margin loss.
Neither scenario is hypothetical in the sense of being invented — the 6.1% figure is StatCan's own reported number for the month. The only variable in the example is how quickly the feed caught up to it.
A single monthly release rarely moves every category the same amount, which is the whole argument for step nine's category-by-category cadence.
The refresh cadence should follow the volatility. The retention clock should not — those are two different rules answering two different questions, even though both start from the same monthly release.
Either a calendar year or a fiscal-period basis is allowed under s. 238(3), but once chosen it cannot be changed without the Minister's authorization — decide it once, deliberately, rather than defaulting to whichever period the software happened to be set up on.
No — route it through the same manual-override log from step four, even under time pressure. A spike is exactly the situation that log exists to catch, not an exception to it.
One quarter is enough to catch drift before it compounds across several quotes. Waiting a full year to reconcile means a stale category like lumber can move several times before anyone checks it against actual invoices.
A 30-minute call is enough to tell you whether it is worth building.
