Anonymised, illustrative composite. A landscaping and civil crew laid off every worker the same way each winter — and lost its strongest people to other employers while it did it, because their paperwork moved slower than everyone else’s.
At a glance
A landscaping and civil contractor laid off its full 22-person field crew every November and recalled as many as it could the following April, the way most seasonal outdoor trades in Canada operate. The prior spring, only eleven of the twenty-two had come back — and the firm’s own performance notes showed the departures skewed toward its most experienced, highest-rated workers, exactly the people it most wanted to keep.
The firm’s first instinct was that this was a wage problem, or a culture problem — that its best workers had simply outgrown seasonal work and moved on to year-round jobs. Exit conversations with a few of the departed workers did not support that story: several said they had assumed they were not being asked back at all, because nobody from the firm had been in touch by the time another employer made them an offer for the winter.
A review of the layoff process found the payroll clerk had been batching all 22 Records of Employment together at the end of the season and processing them over the following two to three weeks, rather than issuing each one as the individual layoff took effect. The Employment Insurance Regulations require the employer’s copy of a Record of Employment to be delivered to the worker “not later than five days after the later of” the day the interruption of earnings occurs or the day the employer becomes aware of it — a Record of Employment is required specifically “in respect of a person…who has an interruption of earnings,” which a seasonal layoff is. That five-day rule governs a paper Record of Employment; for an electronic Record of Employment — the ROE Web format Service Canada now treats as standard — the same regulation instead requires it sent to the Commission not later than the earlier of five days after the end of the pay period in which the interruption fell, or, on a payroll of 13 or fewer pay periods a year, 15 days after the interruption began. Either way, a batch processed 9 to 26 days late blew both versions of the deadline, not just the simpler one. Batch-processing twenty-two of them over three weeks meant the last ones out were issued far outside that window.
The audit found the prior year’s ROEs had gone out between 9 and 26 days after each worker’s last day, against a five-day requirement — every one of the 22 was late, and the workers laid off latest in the batching queue waited longest. A worker whose EI claim cannot start until the Record of Employment arrives has every incentive to take other work in the meantime rather than wait out a paperwork delay through the winter, and the firm’s strongest performers — the ones with the most standing offers elsewhere — were exactly the ones with other options to take.
The firm rebuilt its layoff-day checklist so each worker’s Record of Employment is issued within the five-day window from their actual last day, not batched at season end. It paired that with a written, documented recall order communicated at the point of layoff — explicit priority commitments for its top-rated workers — rather than leaving recall as an undocumented, ad hoc decision made fresh each spring.
The following winter, all 22 Records of Employment went out within the five-day window — zero late, against every one being late the year before. The spring return rate rose from 11 of 22 (50%) to 19 of 22 (86%), with the clearest shift among the firm’s highest performers: of its seven top-rated workers, only three had returned the prior spring; all seven returned this time.
For the underlying mechanism, see how a Record of Employment works for a seasonal layoff, and for the sequencing decision itself, see layoff and recall sequencing and planning a seasonal workforce.
On the firm’s own tracking, replacing a departed crew member — recruiting, screening and getting a new hire productive on a landscaping or grading crew — runs roughly $2,800 per hire. Against the eleven replacements the prior spring’s 50% return rate required, that is about $30,800 in replacement cost the firm was absorbing every single year, on top of losing the specific skill and site knowledge its most experienced workers carried and could not simply be rehired into a new person.
The tell is a return rate that runs backwards — the workers with the most options elsewhere are the ones not coming back, while the ones with fewer alternatives do. That pattern points at something the firm is doing that only its strongest people have the leverage to walk away from, and a slow Record of Employment, which delays the one thing — an EI claim — a laid-off worker is actually waiting on, is exactly that kind of friction.
A 30-minute call is enough to tell you whether AI pays for itself here.
