A retail DC consolidation: mixed SKUs, one pack-out spec

A grocery retailer consolidates fourteen chilled and frozen lines into one pack-out room and asks for a single specification covering all of them. The specification that comes back says “one 800 g gel pack per carton, two for the large format”. It is one page, it is easy to train, and it is wrong for at least four of the fourteen lines.

The failure is not dramatic. It is a pallet of chilled yogurt cups that arrives at 5.9 °C in July when the contract says 4, and a pallet of frozen peas that arrives at minus 12 with a gel pack that did its job too well on the wrong side of the divider. Both were packed to spec. The spec was the problem.

One spec, three thermal jobs

Fourteen SKUs almost always fall into three groups, not one. Group one needs to be held cold: chilled ready meals, dairy, fresh juice. Group two needs to be held frozen and cannot tolerate a thaw: ice cream, frozen pastry, anything where a refreeze shows. Group three is tolerant: hard cheese, some produce, shelf-stable items that are chilled only for the journey.

The packing room does not need fourteen specifications. It needs three, and it needs a rule that assigns every SKU to one of them at the point the order is picked. Anything that cannot be assigned is a product-development question, not a packing question, and it should be flagged rather than defaulted into group one.

The mixed carton is where the money leaks

Consolidation pushes mixed cartons, and mixed cartons are where thermal specifications quietly collapse. A single carton holding frozen peas and a chilled dessert has one air space and one set of packs, and the packs will cool whichever mass is closest to them. The peas win. The dessert sits at 3 °C for six hours, then warms, and the customer finds a soft centre.

The workable answer is not more gel. It is a divider with a real R-value and a rule that no carton mixes group one and group two. Where a mixed carton is unavoidable, the frozen line goes in the middle of the carton with packs on both faces, and the chilled line goes at the end furthest from the packs, against the insulated wall. We have seen this single change cut mixed-carton temperature failures from 7% to under 2% in a retail consolidation, with no increase in pack count.

Building the one-page spec that actually holds

A specification that survives contact with a pick-and-pack operation has four columns and no prose. Column one: group. Column two: pack size and count. Column three: placement, described as a position in the carton rather than as a percentage of volume. Column four: the one thing the packer must not do, written as a single sentence per group.

Keep the fourth column short and physical. “Do not put a group two carton beside a group one carton on the pallet” beats a paragraph about thermal separation, because a packer at 22:00 reads the fourth column and nothing else. Then put a coloured tape stripe on the bench for each group, so the assignment is visible without reading.

Auditing a consolidated room in one shift

Walk the room at three points in a shift and record four things: the temperature of the air at bench height, the number of open master cartons on the floor, the number of cartons whose group tape does not match their contents, and the surface temperature of a sample of ten packs coming out of the freezer. Those four numbers will explain most of the variance your customers are reporting, and they take twenty minutes to collect.

Do it for five shifts before you rewrite anything. Retail consolidation projects fail when the specification is written from a conference room, and they settle down quickly once the numbers from the floor are on the table.

If you are consolidating two pack-out rooms into one this year, the fill weights and placement rules in our SAP powder product range are grouped by hold time rather than by SKU, which maps cleanly onto the three-group structure above.

Pick your four worst-performing SKUs this week, assign them to groups, and run one pallet of each under the new column-three rules. One pallet is enough to show whether the divider is doing its job, and it is a much cheaper test than a full re-specification.