How Many Warehouse Workers Do You Need Per Shift,
and How to Hold That Number

How many warehouse workers you need per shift

Warehouse staffing levels are a number per shift, not a number per building. Each shift owns its own slice of the day’s work, and four figures clear it: the work, the rate, the productive hours, and the jobs that are not picking. Nothing else goes in.

All four sit inside your own operation. Nobody’s software knows your rate. A staffing model built on an industry average is a model built for a building you do not run.

Almost every staffing guide will start you on a benchmark rate. That advice is wrong. A published rate was measured in somebody else’s building, on their best week, with their racking and their pick paths. It will overstate you by fifteen to twenty percent. Measuring your own costs one picker, one shift and a calculator.

The arithmetic is two divisions. Volume divided by rate gives you the labor hours the shift holds. Labor hours divided by productive hours gives you people. Round up, then count the non-picking roles on top.

For example, take the second week of January at a Midwest distribution center, once peak is over and the volume has settled. First shift runs 6:00 AM to 2:00 PM. It carries 4,000 order lines on a weekday.

Your pickers measure 80 lines an hour. After two breaks and a clean-down they get 7 productive hours each. So 4,000 divided by 80 is 50 labor hours of picking, and 50 divided by 7 is 7.1 people. Eight pickers. Two forklift operators have to be on trucks the whole shift. First shift is ten people.

Round up, always. A picker with a quiet half hour is cheaper than a truck that leaves half loaded. Round 7.1 down and you have 560 lines still on the floor at two o’clock, with a carrier cut-off that does not move for you.

All of this gets thinner as the crew gets smaller. Somewhere under fifteen people the arithmetic stops helping you much, because one absence is seven percent of your floor and no buffer figure survives that.

Going the other way costs real money. It never shows as a line item. One picker more than you need, on five first shifts a week at $22 an hour, is $880 a week. That is about $45,760 a year, out of one site.

That is the whole of warehouse labor planning. Get the number right. Then make sure the number is what actually gets built. I would never set a staffing level off the headcount you remember running last month. Measure the rate.

How to set warehouse minimum staffing requirements per role

A staffing minimum is the smallest number of people, in a named role, that a shift is allowed to be built with. In XShift AI you set one per site and per role, and every shift it builds at that site is sized from it. Not warned about. Sized from it.

Where a shift needs a different number, an override replaces the site figure for the days and the hours you name. That distribution center needs three numbers, not one. Eight pickers and two forklift operators as the floor. Ten pickers on a Monday. Two pickers on third shift.

Each of the last two replaces the floor for the days and the hours you give it, so the rest of the week keeps its original number.

1. 1

   Name every job somebody has to be doing on a shift: Picker, Forklift Operator, Packer, Loader, Replenishment. A count can then be held against each one separately.

2. 2

   Set the floor for the site. Eight pickers and two forklift operators. Ten people, and nothing at that site gets built below it.

3. 3

   Set the Monday figure at ten pickers, for the hours first shift actually runs. It replaces the floor on those hours only.

4. 4

   Set the third shift figure at two pickers, for the hours third shift runs, so the night is sized off its own volume rather than the day’s.

5. 5

   Build the week. Each shift is sized from whichever number matches its site, its day and its hours, and each role is filled separately until that role’s count is met.

In XShift AI you set those three numbers once and every week it builds is sized from them. The screen-by-screen version is on the role-based staffing page.

I would keep both, every time. A warehouse almost always needs both. You want a floor that never moves, and you want the two or three shifts a week that are nothing like the others to carry their own number.

Setting your first warehouse staffing minimums

You do not have to work any of this out before you start. During setup the AI Copilot asks how you staff: by site, or by day and time. It writes your first set of minimums from what you tell it, and the numbers you hand it are the ones the arithmetic above produced. Nothing in there is a guess.

Change them whenever your volume changes, and the change applies to the next week you build.

Every slot it could not fill, and the reason behind it

A minimum is only worth having if you find out the moment it is not met. When XShift AI builds a week, it records every slot it could not fill. It groups those slots into the handful of causes behind them, ordered by how many slots each cause accounts for.

That is the Schedule Log. You read it straight after a week is generated. It counts in slots, not shifts. A shift needing eight pickers with six on it is two people still needed, and that is the difference between a week that runs and a week that limps.

For example, here is what the run above reports back on a bad week.

Fourteen missing people. Three sentences explain all of them. Eight of the fourteen are one problem with one fix. People work at that site, they just do not hold the Forklift Operator role, so give two of them the role and eight slots close at once.

That is why the causes are ordered by size: the row at the top is the change that clears the most work.

Open any cause and you get the shifts underneath it, and under each shift the people who could have worked it, every reason they could not, closest to allowed first.

A week where nobody ever asks why is a week where the same two roles come up short every Monday and nobody connects the two. I read the log before the calendar.

One new shift, start to finish

The whole thing runs in a few minutes, from the volume on the report to the week on the board. Here it is once more, on a different site, following the same six steps. Same order, different numbers.

For example, take the week a customer adds Saturday outbound at a cold storage site. That Saturday carries 1,200 lines.

Your pickers are slower in the cold, at 60 lines an hour. People blame the temperature for that, though I would put more of it on the gloves. With the extra breaks the cold room needs, they get 6.5 productive hours on an 8:00 AM to 4:00 PM shift.

So 1,200 divided by 60 is 20 labor hours, and 20 divided by 6.5 is 3.1 people. Four pickers, plus one loader on the dock.

So the number you write down is four pickers and one loader, at the cold storage site, on a Saturday, from eight to four. Loader has to exist as a role of its own before it can carry a count, because a count on Picker will never put a loader on the dock.

Then generate next week. The Saturday shift comes out sized at five, four pickers and one loader, filled from the people who work at that site and hold those roles. The weekday shifts keep the numbers they already had, because the override only claims Saturday from eight to four.

There is one gap worth knowing about. When you type a shift in by hand and put a name on it, your written rules still run, but availability is not checked on that path. Look at the person’s availability before you save.

Warehouse staffing levels: your questions answered

Related articles