Packing KPIs to track every week: orders per hour, accuracy and more

I cannot give you a good orders-per-hour number.

Anyone who quotes one figure for every D2C floor is guessing. A mailer with one T-shirt and a carton with three glass jars are different jobs, and they will never pack at the same speed.

So instead, here are the numbers we would put on a wall every week, how to work out each one, and what each one hides.

There are six. They cover speed, accuracy, damage, consumables, cut-off and backlog. Any one of them read alone will mislead you, so read them together.

Why weekly, and why so few?

A day is too noisy. One late truck or one big influencer order and the numbers jump. A month is too slow, because by the time you see a problem, four weeks of orders have already gone out with it.

A week is long enough to show a trend and short enough to act on.

And six is enough. A board with fifteen numbers gets read by nobody.

I learnt the value of small, steady improvement from our own tape plant. Our plant head worked on scrap quietly, one small change at a time, trying to be 1% better every day. I had read Atomic Habits long before, but I had not used it on the shop floor. He did. A weekly board is the same idea for a packing floor.

1. Orders per packer-hour

Orders packed, divided by the hours packers actually spent on packing stations.

Use station hours, not shift hours.

If a packer spends an hour unloading a truck, that hour does not belong in the packing number. Otherwise you are measuring your rota, not your packing.

Let us take a worked example. The numbers are assumptions, for illustration only. Say 4 packers spend 7 hours each on station on Monday, and they pack 1,120 orders. That is 1,120 ÷ 28 = 40 orders per packer-hour.

Split it by order type. Single-item and multi-item orders should be two lines on the board, and so should mailers and cartons if you pack both. A change in your order mix can move the total without anyone packing faster or slower.

We would take the order count from the label print log or the last scan, and the hours from your roster. To find out where the seconds go inside an order, use the video method in designing a fast packing station.

What it hides: speed bought with errors and damage.

Never celebrate this number without looking at the next two.

2. Packing accuracy

Mis-shipments that started at packing, per 1,000 orders packed. Put the total for the whole floor (pick, pack, data and dispatch) on the line below it.

The source is your root-cause log. If you do not keep one yet, how to cut mis-shipments shows the columns we suggest.

Also count the errors your team caught before dispatch. A rising count of caught errors is good news. It means the checks work and people are reporting.

What it hides: errors customers never report. A customer who receives a wrong variant may simply keep it, or quietly stop ordering.

3. Damage rate

Damaged deliveries per 1,000 orders delivered, counted from complaints with photos, returns marked damaged, and replacements sent.

Report it by dispatch week, not complaint week. Damage is reported days or weeks after the parcel left, so a complaint that arrives this Friday belongs to an earlier week's packing.

Treat it as one input, not a packing score. The product, the pack design, the packer's method and the courier's handling all feed into it, and a damage rate that jumps in the same week you add a new courier says more about that lane than about your packers. Split it by SKU and by courier before you blame anyone. If one courier's lane shows most of the damage, the packing station is not the first place to look.

4. Consumables per order

Tape, void fill, boxes and mailers issued from the store, divided by orders packed. Compare each with your packing standard, i.e. what the specification says one order should use.

Tape is the easy one to measure. Rolls issued, multiplied by metres per roll, divided by orders. Say 60 rolls of 65 m went out in a week of 3,000 orders (again, assumed numbers). That is 3,900 m ÷ 3,000, or 1.3 m of tape per order. If your standard says 1 m, someone is adding a strip.

That gap is worth a conversation.

Because we make tape and void fill, this is the number I look at first. A packer who adds an extra strip "to be safe" on every carton is being careful, not careless. The fix is a clear standard and a tape that holds the first time, not a lecture.

Add one more line here: box mix, the share of orders packed in the box size the specification asks for. A bigger box than needed means more void fill, more tape and a heavier parcel on the courier's scale.

What it hides: under-use. A number falling below standard can mean a saving, or it can mean parcels going out under-protected. Check it against the damage line.

5. Cut-off adherence

Orders received before your dispatch cut-off that were handed to the courier the same day, divided by all orders received before cut-off.

We would write down the reason for every miss: stock not found, address problem, late pick, packing backlog, courier late. After four weeks, the reasons tell you where to work.

A second check sits beside it. When the courier arrives, are the parcels already sorted and manifested? A pickup that waits while the floor scrambles is a cut-off you nearly missed.

6. Backlog at close

Orders received but not packed when the floor closes, and the age of the oldest one in hours.

This is the early warning. A backlog that grows a little every day for a week is a capacity problem forming quietly, long before a customer complains.

What does the weekly board look like?

One page, on a wall near the packing area, or one screen. This is the layout we suggest:

We suggest setting each target from your own 4-week baseline, then asking for a small step at a time. Copying a number from another company's floor sets your team up to fail or to coast.

Our advice is to run the review in 20 minutes, on the same day every week, standing at the board. Bring a packer and a picker along with the supervisors. Every number in red gets one action, one owner's name and a date. If the same number is red three weeks running, the action was the wrong one.

Which mistakes make a KPI board useless?

1. Speed with no accuracy beside it.
2. Shift hours instead of station hours.
3. Comparing packers who pack different kinds of orders.
4. Numbers filled in from memory instead of system data.
5. Adding a seventh and eighth number before the first six are steady.
6. Using the board to blame people, which teaches them to hide errors.

The same board works when you are testing a new courier or fulfilment partner. Run it on the trial lanes, as described in choosing a 3PL or courier partner.

Put the six numbers on a whiteboard this Monday, even if half the boxes are blank. A blank square on a wall gets filled faster than a column nobody opens. Write to us through the contact page and tell me which number surprised you.