GC closes out a school in nineteen days

The situation

On the GC’s prior comparable institutional job, the building was physically finished well ahead of its own closeout package. Product data sheets, shop drawings, and approval records for the mechanical and electrical scope existed — they were just scattered across a dozen subtrades’ email threads, none of it centralized until someone went looking for it after the fact. Assembling the accepted closeout package from that starting point took 11 weeks.

The school addition had no such runway. The board’s occupancy date was fixed months in advance and tied to the new school year; there was no version of closeout running a few weeks past substantial completion that worked here.

Institutional and public-sector work carries this pressure more than most private commercial jobs, precisely because the occupancy date is set by someone else’s calendar — a school board’s term dates, a hospital’s unit transfer schedule, a government tenant’s lease start — and it does not move to accommodate a contractor’s closeout process. A GC that treats closeout as a construction-phase afterthought can usually absorb the slippage on a job with a negotiable handover date. On a job like this one, there was nothing to absorb it with.

The problem

The fix the GC made was not a new closeout procedure — it was moving an existing tool earlier in the schedule. Procore’s Submittals tool lets a team “organize and manage their submittal log by division, attach shop drawings and product catalog pages, and track approval workflows,” and supports building submittal packages that bundle related items together for consolidated review and distribution. None of that is closeout-specific functionality — it is the same submittal log every subtrade’s product data already runs through during construction. The only change was treating it as the closeout package from day one, rather than a construction-phase tool that closeout would draw on later.

The numbers

From mobilization, every subtrade’s submittals were logged against their CSI division in the register as they were approved — not re-requested at the end. By the time substantial performance was reached, the bulk of what a closeout package needs was already sitting in an organized, approved state inside the same log the project had been running all along. What remained after substantial completion was assembling the binder and clearing the last handful of items, not chasing subtrades for documentation they had already submitted months earlier.

The prior comparable job’s 77 days fell to 19 — a 75% reduction (77 minus 19 is 58; 58 divided by 77 is just under 75%). The board took occupancy on its target date.

The rule that decided it

The occupancy date does not move for any GC on any job — it is fixed well before ground is broken. The only variable a GC actually controls is how much of the closeout package already exists by the time that date arrives. A submittal log run by division, with an approval workflow attached, does that continuously through construction if it is used that way; treated as a construction-phase administrative task instead, the same log still exists at the end, but nothing in it was organized around closeout, and someone has to do that organizing after the deadline has already started counting down.

Ontario’s Construction Act gives “substantial performance” itself a precise numeric test, not a percentage-complete estimate: a contract is substantially performed once the improvement is ready for use and the cost to complete or correct any known defect is no more than 3% of the first $1,000,000 of the contract price, 2% of the next $1,000,000, and 1% of the balance (s. 2(1)). That test, not a site walk-through, is the clock this GC was actually racing — see how a Construction Act holdback is calculated and released.

The outcome

The school opened on the board’s target date. The GC now runs every institutional project’s submittal log by division from mobilization, treating closeout as something the log builds continuously rather than something assembled at the end. See a related change to how deficiencies get walked, and for the underlying term see what a submittal register actually tracks.

For how firms are applying AI tools to the submittal review workload itself, see AI-assisted submittal review.

Takeaways

• • Closeout speed is set by how much of the package already exists at substantial completion, not by how hard anyone works after that date.
• • A submittal log organized by division is a construction-phase tool by default — it becomes a closeout tool only if it is treated as one from mobilization.
• • A fixed occupancy date, common on institutional and public-sector work, removes the option of letting closeout run long; plan the submittal log around that date from day one.
• • Measure closeout time against your own prior comparable project. A generic industry figure does not tell a GC what its own process is capable of.
• • Public-sector occupancy dates are set by someone else’s calendar, not the contractor’s — a school term, a lease start, a unit transfer — and do not move for a slow closeout.
• • Bulk-review and package-distribution features in a submittal tool matter most in the last two weeks before handover, when a GC is assembling a binder rather than chasing individual documents.