Choosing pack sizes for a mixed-order week

Five order profiles in one week: fourteen 2 kg boxes on Monday, sixty 8 kg on Tuesday, twenty-two at 15 kg on Wednesday, nine at 22 kg and two at 30 kg on Thursday, then a repeat of Monday on Friday. The plant stocks three pack sizes because that is what the freezer holds, and the packer on Wednesday spends more time deciding which combination to use than building the carton. Fill time per box: four minutes forty. Target: two minutes.

Pack size selection is usually argued as a thermal question. On a mixed-order week it is an operational one, and the operational answer usually improves the thermal result as a side effect, because a simpler build has fewer ways to go wrong.

Start from the order histogram, not the product list

Pull ninety days of orders and plot box mass against order count. Most businesses find two or three peaks and a long tail, and the peaks are where the money is. If 70% of orders fall between 6 and 10 kg, a pack sized for 8 kg with a small trim pack for the edges will cover most of the week, and a single large pack handles the tail.

What you are looking for is the smallest number of pack sizes that covers 85% of orders without a combination. Combinations are where the time goes and where the errors live.

The two-size-plus-one-sheet plan

A workable structure for most mixed weeks: one small pack for orders under 6 kg, one large pack for 6 to 18 kg, and a single refrigerant sheet that can be added to either for the 18 to 30 kg tail. Three components, one decision rule, and a build that a seasonal worker can learn in an afternoon.

The sheet is the important part. It gives you a way to extend the large configuration without introducing a third pack size, and it stacks flat in the freezer so it costs almost nothing in storage. Where the tail is very small, the sheet can be replaced with a second pack of the small size, which the plant already stocks.

What the numbers look like

In a plant running the arrangement above on the profile at the top of this article, average build time fell from four minutes forty to two minutes twenty, and pack-related errors fell by roughly two-thirds, because the decision reduced to one question about mass. The freezer held two pack SKUs and one sheet instead of three pack SKUs, which freed about 12% of the frozen volume.

The thermal result improved too, though not for the reason people expect. It was not that the new configuration was better in a chamber. It was that the packer stopped making the wrong choice under time pressure, and the wrong choice was previously responsible for most of the out-of-range shipments.

Where selection goes wrong

The classic error is adding a pack size to solve a thermal problem that is really a placement problem. A pack in the wrong position can underperform a smaller pack in the right one, and adding a size does not fix placement. Before you add a SKU, run a placement test on the existing configuration; if moving one pack recovers the hold time, you have saved yourself a freezer slot and a training problem.

The second error is sizing for the worst order and using it for everything. A 30 kg configuration on a 2 kg order wastes gel, freezer capacity and money, and it produces the “too much packaging” complaint that erodes trust faster than a warm box.

The third error is subtler: sizing packs to the box rather than to the payload inside it. Two orders can arrive in an identical carton with 3 kg of dense product in one and 3 kg of bulky salad in the other, and the air space is what decides the pack requirement, not the mass. Where the mix includes both, put an air-volume column in the selection table and let the packer read that column, because box volume is the number that bites in a week with wide product variety.

If you are rationalising a pack range, the size and hold-time matrix in our SAP powder product range is organised so that one grade covers a family of sizes, which keeps the SKU count down without a separate qualification per size.

Pull ninety days of order data this week and count the peaks. The number of pack sizes you actually need is almost always smaller than the number you stock.