The delay claim that survives isn't the one with the best story afterward. It's the one with a Delay Entry logged the day the crew actually stood down, against a baseline that was set before the job started.
Key takeaways
A delay claim is won or lost on documentation created while the delay was happening, not on reasoning constructed after the fact to explain why the schedule slipped. The two things that actually make a claim defensible — a same-day record of what stopped work, and an objective baseline to measure it against — both have to exist before the claim is ever written up, not assembled around it afterward.
Procore's Daily Log tool includes a dedicated Delay entry type among the field categories it tracks, alongside Weather, Manpower, Equipment and Productivity entries — and logging a delay the same day it happens, in that same record, is what turns “the job ran late” into a claim with a specific, contemporaneous date, cause and crew impact attached to it. A delay claim assembled months later from memory and a project schedule alone is reconstructing exactly the kind of record that should have existed from day one — and reconstruction is always weaker evidence than something logged the same afternoon it happened.
“It rained more than usual” isn't evidence on its own — a weather-delay claim needs to be measured against a published, objective baseline for the site's region. Environment and Climate Change Canada's climate normals provide exactly that kind of yardstick: its 1991–2020 normals for Kelowna, for instance, record a median frost-free period of 130 days and a 90th-percentile frost-free period of 156 days for that station. The specific figures differ by region and station, but the mechanism is the same everywhere: pull the normals for the actual project site, and measure the days actually lost against what that baseline says is typical for the season, rather than asserting a day was unusual without anything to compare it to.
Worked example: pricing a documented weather delay
Two frost events in April stop exterior work on days the site's own seasonal baseline would normally have supported it, logged as Delay entries the same day each occurred. Across both events, a five-person crew loses six full working days.
Direct labour cost of the idle crew: 5 workers × $62.00/hour × 8 hours/day × 6 days — $14,880.00.
Extended general and administrative cost at a stated daily rate of $350.00 across the same 6 days adds $2,100.00. Total documented claim: $16,980.00 — a figure built entirely from logged days and a stated crew rate, not an estimate of the schedule's overall slip.
A revised schedule showing a later finish date doesn't, by itself, prove what caused the delay or when the owner or contract administrator was made aware of it. CCDC's own guide to contract administration — CCDC 20, addressing the practices involved in administering a construction contract, intended primarily for the contract administrator — frames written correspondence as the working record of how a contract is actually administered day to day. A delay or constructive-acceleration claim is built on exactly that correspondence trail: the notice given when the delay started, any response received, and the record of whether the contract administrator directed acceleration to recover the lost time. Without that written trail, a claim is left arguing from the schedule alone, which shows the effect of a delay but not who caused it or when it was raised.
A delay only actually pushes the finish date if it hits an activity on the schedule's critical path, or eats through the float on an activity that isn't yet critical. Logging a delay entry without also noting which schedule activity it affected, and whether that activity had float remaining, leaves a real gap in the claim — a delayed activity with three weeks of float absorbed the lost time without moving the finish date at all, while the same delay on a zero-float critical activity pushes the whole project. A rolling three-week lookahead reviewed against the log is what catches that distinction while there's still time to recover the float on a near-critical activity, rather than discovering after the fact that an activity everyone assumed had slack actually didn't.
A single Delay entry on one afternoon is rarely enough on its own to support a meaningful claim. What actually carries a delay or acceleration claim is the same discipline applied consistently across every affected day — logged the day it happened, tied to the site's own baseline where relevant, and backed by the correspondence trail showing the impact was raised at the time. The same daily discipline that supports a delay claim is also exactly what an early warning that a job is going sideways is built from — a field record kept consistently, not reconstructed under pressure. Where the delay also affects the price of the work, the same itemized pricing discipline applies to the acceleration cost as to any other change — labour, equipment and any premium cost incurred to recover the schedule, each traceable back to a specific logged day rather than a single lump-sum figure attached to the claim at the end.
That documentation discipline reaches further than the immediate dispute. Ontario's Construction Act defines the “price” a lien secures to include, under s.1(1)(b), “any direct costs incurred as a result of an extension of the duration of the supply of services or materials to the improvement for which the contractor or subcontractor… is not responsible” — delay costs, in other words, aren't just a contractual argument, they're themselves part of what a properly documented lien can register to protect.
Immediately, the same day it happens, as a Delay entry in the daily record. The schedule impact can be assessed later; the contemporaneous record of what happened and when has to exist from the day of the event, not be reconstructed afterward.
Measuring the days actually lost against an objective, published baseline for the site's own region — such as Environment and Climate Change Canada's climate normals — rather than simply asserting a particular period was unusually bad.
Not on its own. A schedule shows the effect of a delay but not who caused it or when it was raised. The written correspondence trail — notice given, response received, any acceleration directed — is what actually supports the claim.
Rarely by itself. What carries a claim is the same documentation discipline applied consistently across every affected day, not a single logged event treated in isolation.
A delay on an activity with float remaining may not push the finish date at all, while the same delay on a zero-float critical activity pushes the whole project. Logging the delay without noting which activity it hit, and whether that activity had float, leaves that distinction out of the claim.
A 30-minute call is enough to tell you whether AI pays for itself here.
