Pick, pack, dispatch: how to cut mis-shipments

Two rolls of packing tape that differ by a few microns look exactly the same. So do two cartons of the same tape in different widths, until you read the small print on the side.

We make these products in Jaipur, so I know how alike they look on a shelf. D2C warehouses face the same thing every day with sizes, colours and pack counts.

In my view, most mis-shipments are not careless people. They are a process that lets a look-alike through.

The direct answer is a check at every handover. Scan at pick, scan every unit at pack, check the weight, photograph the open box, and scan every parcel onto the courier's manifest. Then log each error that still gets through, and fix the cause rather than the person.

What counts as a mis-shipment?

We suggest agreeing on this list first, or every team will count differently:

1. Wrong item. A different product altogether.
2. Wrong variant. Right product, wrong size, colour or pack count.
3. Wrong quantity. One short, or one extra.
4. Missing item in a multi-item order.
5. Swapped parcel. The right contents with another customer's label.
6. Not handed over. Packed, but it never left with the courier.

I would count them per 1,000 orders shipped. The sources are customer complaints, return reason codes, courier exceptions, and errors your own team catches before dispatch.

I like to keep those internal catches as a separate count.

They are near misses, and they show which checks are working.

Where do mis-shipments start?

Every error starts in one of four places: data, pick, pack or dispatch.

The first one surprises people. If a marketplace listing is linked to the wrong SKU in your system, the floor will ship the wrong item perfectly, every single time. No floor check catches it, because the screen and the item agree with each other.

So before you blame the pickers, I would check the mapping. Take your top 50 SKUs and confirm that each channel listing points to the right internal SKU and the right barcode.

How do you cut errors at picking?

1. Print or show pick lists in walking order, with the SKU, the variant and the quantity in large type, plus a product photo.
2. Keep one SKU per bin. Never mix two SKUs in one location, even for a day.
3. Keep look-alikes apart. Two sizes of the same product in neighbouring bins invite the wrong pick. Separate them, or mark each bin with a large, contrasting size label.
4. Scan the bin, then the item. The system checks both against the pick list.
5. Scan every unit. The jugaad of scanning one unit and typing "3" is exactly where quantity errors live.
6. Never substitute. If the bin is short, the order goes to the exceptions tray with a note, not out with a "close enough" item.

We would also insist that every variant has its own barcode. If a supplier prints one code on all sizes, put your own SKU label on each unit at receiving, before it reaches a shelf.

Slotting decides how many look-alikes sit together in the first place, and our guide to warehouse layout and slotting covers it.

What should the packing station check?

Packing is the last moment the box is open. If you can afford only one check, our advice is to put it here.

1. The screen shows each item's photo, variant and quantity in large type, so the packer compares the item in hand with the picture.
2. Every unit is scanned again, and the shipping label prints only when the scans match the order. I rate this as the strongest single check, because nothing can change after it.
3. One order is open at a time. The first order is sealed and labelled before the next tote is opened. Two open orders on one bench is exactly how labels get swapped.
4. The label goes on the parcel the moment it prints, on the parcel in front of the packer and nowhere else.

The station layout that makes this easy is in designing a fast packing station.

Is a weight check worth adding?

Yes, for most product ranges. Weigh each sealed parcel and compare it with the expected weight: the items plus the packaging.

We would set the tolerance from your own parcels. Weigh 20 correct parcels of a typical order and look at the spread. A parcel outside that spread goes to the exceptions tray and is opened.

I should be honest about its limit. A weight check works when items differ in weight. A missing 20 g lipstick in a 2 kg parcel will sit inside any sensible tolerance. For light add-ons in heavy orders, the scan is your only real check.

Is photo proof worth it?

A photo of the open box, with the items visible and the order number in the frame, takes seconds. A camera fixed above the station and triggered by the last scan takes no packer time at all.

Let us be clear about what it does.

A photo does not stop an error. It records one.

That record still pays for itself. It settles disputes with customers and channels. It also shows a trainer, or the root-cause log, exactly what went into the box. Point the camera at the box, not at the packer's face, and keep the photos for as long as your returns and dispute window runs.

What checks belong at dispatch?

1. Sort parcels by courier into marked cages or sacks.
2. Scan every parcel at handover against the courier's manifest, and count the parcels on both sides before anyone signs.
3. Any parcel that fails to scan or is not on the manifest goes to a marked exceptions spot, never back into a random pile.
4. Before you close each day, packed-but-not-handed-over should be zero. Anything left gets a written reason.

This is where "not handed over" errors are caught.

They are the easiest to miss, because the customer finds out first.

How do you run a root-cause log?

We suggest one sheet, with one row per mis-shipment, with these columns: date, order number, error type, where it started (data, pick, pack or dispatch), what happened, why it happened, the fix, the owner and the date it was closed.

For the "why", keep asking until you reach a process cause. Let us take an example to show the method.

1. Why did the customer get a large instead of a medium? The picker took it from the neighbouring bin.
2. Why was that possible? The medium and large bins sit side by side with identical photos.
3. Why are they side by side? Slotting grouped the product family together.
4. The fix: separate the sizes, and add large size labels on each bin.

I would review the log for 20 minutes every week, with the people who pick and pack. A single slip by a new packer is a training point. The same error twice, in different hands, is a process problem.

When a packer catches an error at the bench, I would say so in that meeting, by name. A caught error is a check doing its job, and people report more when reporting is thanked rather than punished.

Your weekly accuracy number belongs on the same board as speed, and packing KPIs to track every week shows how we would set that board up.

Start with the mapping check on your top 50 SKUs this week. It is a small job, and it is the one error that no amount of care on the floor can fix.