Warehouse turnover: why it happens and what actually fixes it

Ask a warehouse manager why they can't keep people and you will almost always hear a version of "nobody wants to work" or "they leave for fifty cents more down the road." Both are occasionally true. Neither explains the pattern we actually see, which is much more specific and much more fixable.

The pattern is this: most warehouse departures happen in the first two weeks, and almost none of them are really about pay.

1. The shape of the problem

Warehouse and light industrial turnover is not spread evenly across a worker's tenure. It clusters hard at the front:

• Day one to day three. The largest cluster. The job was not what was described, nobody was expecting them, the physical demand was a shock, or the commute turned out to be impossible. Almost entirely a communication and first-morning problem.
• Week two to week four. The second cluster. The shift pattern collides with the rest of their life — childcare, a second job, transit that doesn't run at the hour their shift ends. Usually a scheduling problem.
• Month three to month six. A smaller, different cluster. They've learned the job, they're good at it, and they can now see there is nowhere to go. This is the one that costs you most, because you trained them.
• After a year. Genuinely low, in most operations. People who make it past six months tend to stay.

That distribution matters because it tells you where to spend. Raising the rate does very little for the day-one cluster — the person who walked off at 10 a.m. did not do a wage comparison. Most retention budgets are aimed at the smallest cluster.

2. The seven real causes

In rough order of how often we see them:

The job wasn't what was described

"Some lifting" turns out to be the entire shift. "Warehouse work" turns out to be a freezer. The posting said picking; the reality is unloading containers by hand. This is the single biggest cause and it is entirely self-inflicted — a vague job description filters nobody out, so it fills your first day with people who were never going to stay.

The commute didn't survive contact with reality

A site fifteen minutes away by car is a ninety-minute two-transfer transit trip that doesn't run before 5:30 a.m. or after 11 p.m. The worker does it twice on goodwill and then stops. This is the most predictable cause on the list and almost nobody checks it before the offer.

The first morning was chaos

They arrive at 6 a.m., the supervisor who hired them is in at 8, nobody at the door knows who they are, and they stand around for forty minutes. Some leave before anyone talks to them — and get recorded as a no-show. That first hour sets whether the person believes this is a serious place to work.

The physical reality was a shock

Eight hours on concrete, in a cold room, at pace, is genuinely hard, and people who have never done it don't know what it costs until day two, when everything hurts. Preparing someone for this honestly — and easing the ramp in the first three days — retains people that surprise loses.

Rate or hours ambiguity

They understood one number, the first pay shows another. The premium they thought applied doesn't. They were told "40 hours" and got 26. Trust does not recover from this, and it should not.

Supervisor churn and inconsistency

Three different people telling them three different methods in the first week, or one supervisor whose approach is corrective only. People do not leave warehouses; they leave the specific person standing over them. Where one team's turnover is triple another's on the same site, the difference is not the work.

Visibly nowhere to go

The month-three cluster. Nobody ever mentioned certification, lead hand, a different department, or a permanent conversion. A competent worker who can see no path takes the next fifty cents, because there's no reason not to.

3. The first ninety minutes

If you fix one thing, fix this. It costs nothing and it moves the biggest cluster.

1. Someone by name is expecting themNot "reception," not "ask for the supervisor." A named person who knows the new worker is coming, at the actual start time, and who is present. If the shift starts at 6, that person is there at 6.
2. Ten minutes of orientation before any workWhere the bathrooms are, where breaks happen, where to put their bag, what the break times are, who to ask when something goes wrong. Trivial for you; the difference between belonging and being a stranger for them.
3. Safety walk-through, done properlyHazards, equipment, emergency exits, lockout, what to do if something is unsafe. This is also your legal obligation as the party controlling the workplace — it does not transfer to the agency.
4. One named person to ask questions of, all weekNot "anyone." A buddy. New people won't interrupt a stranger to ask a basic question, so they guess — and guessing is how errors and injuries happen.
5. A check-in at the end of day oneTwo minutes. "How was it, is anything unclear, will we see you tomorrow." You will hear the reason someone was about to disappear, while it's still fixable.

Almost every first-week departure we investigate was visible on day one and nobody asked a question.

4. The commute arithmetic nobody does

Take the site postal code, take the shift start time, and check the real transit journey at that hour — not the midday route. Then ask whether you would personally do that trip twice a day for the rate you're offering.

This is not a small factor in our markets. A distribution site in the East End of Montréal, in Burnside in Dartmouth, or in an industrial park outside Moncton can be effectively unreachable at 5:30 a.m. by transit, which quietly restricts your entire candidate pool to people with cars — while your rate assumes people without them.

What actually helps, in order of cost: recruit deliberately from the transit lines and neighbourhoods that do reach you; align shift start with the first realistic bus rather than five minutes before it; where several people share a route, coordinate a carpool; and if you're on a genuinely unreachable site with an early start, accept that a transport allowance is cheaper than replacing the position four times a year.

5. Shift patterns that break people

Some patterns lose people regardless of pay:

• Schedules published with two or three days' notice. This makes childcare impossible and a second job impossible. It is the most common and most damaging pattern we see. Two weeks of visibility retains people at the same wage.
• Rotating shifts with a fast turnaround. Finishing at 11 p.m. and starting at 7 a.m. is a fatigue and error generator before it's a retention problem.
• Unpredictable mandatory overtime. Being told at 2 p.m. that you're staying until 8 works once. As a pattern it costs you your most reliable people first, because they're the ones with the most other commitments.
• Chronically short hours. Somebody promised full time and averaging 26 hours will take any other offer, and should.

6. What actually retains — cheapest first

1. An honest, specific job description. Free. Real lift weights, real temperature, real pace, real shift. Fewer applicants, dramatically better first-week survival.
2. A named person for the first morning. Free. Biggest single lever on the day-one cluster.
3. Two weeks of schedule visibility. Usually free — it's a planning discipline, not a cost.
4. A day-two conversation. Two minutes of a supervisor's time.
5. Naming a path out loud in week one. Nearly free. "Six months solid and we'll certify you on the reach truck" changes how someone weighs a competing offer.
6. Consistent supervision. Costs training attention, not payroll. The highest-return management investment on most floors.
7. Small physical fixes. Anti-fatigue mats, decent gloves, a warm break area for cold-room crews, boots that fit. Cheap, and read as respect.
8. Rate. Last, deliberately. It matters — and it is the only item on this list that recurs every hour of every week. Spend the free items first, then pay competitively for the people who stayed.

7. Three numbers to measure

You cannot manage this on impressions. Track three things, by position and by supervisor:

• First-week survival. Of everyone who started, what share worked a second week? This is your job-description and first-morning score.
• 90-day survival. The headline number. Under 60% means the problem is structural, not the labour market.
• Turnover by supervisor. Uncomfortable and by far the most informative. When one team on the same site loses three times as many people as another, you have found your actual cause.

8. Where an agency helps — and where it doesn't

We will happily fill a position four times a year. We would rather tell you why you're filling it four times.

An agency genuinely helps with the sourcing side: a screened pool, honest briefing of workers on the real physical demand before they arrive, and a day-two call that surfaces problems while they're still small. Where we place repeatedly on the same site, the retention difference usually comes from that call, not from better candidates.

What no agency can fix: a shift pattern that makes childcare impossible, a site nobody can reach at 5:30 a.m., a supervisor whose team always empties, or a job description that misrepresents the work. Those are yours. If turnover is your real problem, the sourcing spend is a tax on not fixing them — and the arithmetic of that tax is in the cost of a bad hire.

---

Related: how temporary staffing actually works — particularly the section on what a strong brief contains, since half the causes above are briefing failures. If your volume is seasonal, the seasonal hiring calendar covers planning the peaks so you're not hiring in a panic.