Mix-up analysis: why discrepancies cancel each other

After a count, the overall difference can be almost zero and yet the warehouse is still in a mess. The reason is a mix-up: one item comes out short while a similar one comes out over, and the two differences cancel each other. Looking at the total, it is easy to think everything is fine. In fact this is the most useful signal of where the errors come from.

Net and gross difference

Calculate two figures together. The net difference is the sum of all differences with their signs (surplus and shortage cancel out). The gross difference is the sum of the absolute values, ignoring the sign. When the net difference is far smaller than the gross, there are many mix-ups.

For example, take a group of phone accessories (records and actual count, units):

• USB-C cable 1 m: 120 and 105, difference −15, price 25 000 so'm, so −375 000.
• USB-C cable 2 m: 80 and 92, difference +12, price 35 000, so +420 000.
• Charger 20W white: 60 and 48, difference −12, price 90 000, so −1 080 000.
• Charger 20W black: 40 and 51, difference +11, price 90 000, so +990 000.
• Earphones B: 30 and 29, difference −1, price 70 000, so −70 000.
• Phone case Y: 90 and 94, difference +4, price 30 000, so +120 000.

Two more items show no difference. The net difference in value is −375 000 + 420 000 − 1 080 000 + 990 000 − 70 000 + 120 000 = +5 000 so'm, and −1 in units. The gross difference, however, is 3 055 000 so'm and 55 units. The report says "almost no differences", yet in fact 55 units do not match the records.

Finding the pairs

Add up the differences by product group (cables, chargers) and look within a group for items with opposite signs and similar quantities. In our example there is a pair of cables (−15 and +12) and a pair of chargers (−12 and +11). The paired quantity is 12 + 11 = 23 units, or 46 units counting both sides. The remaining 55 − 46 = 9 units are genuine differences: 3 + 1 + 1 short and 4 over. Once a pair is found, check physically first: recount both items.

The mix-up share by value is (300 000 + 420 000 + 990 000 + 990 000) ÷ 3 055 000 = 2 700 000 ÷ 3 055 000 ≈ 88 percent. When prices differ, a mix-up affects value as well: the records show a cheap cable, but a dear one is lying there, so for the 12 units the records are understated by 12 × 10 000 = 120 000 so'm.

Causes

Mix-ups are especially common in goods that differ by size, dose, grade or colour: clothing, medicines, building materials, car parts. Typical causes:

• Similar goods stand side by side and differ only in colour, size or model.
• One barcode or one record is used for several variants (colour, size).
• The wrong variant is scanned at receiving, or picking is done by name rather than by code.
• Units of measure get mixed up: a carton or a set instead of a piece.
• In online sales, the product variant is linked incorrectly to the listing on the site.

Fixing it

Do not correct the records by the net figure: if you "close" surplus with shortage, the trace of the error disappears and it will recur. Draw up a separate mix-up document for each pair. Then change the process:

1. Place look-alike goods farther apart.
2. Write the code and colour large on the label.
3. Confirm with a scanner at picking and packing.
4. Open a separate record for each variant.
5. Check groups with a high mix-up share more often by cycle counting.

Conclusion

The net difference hides the consequences of an error, while the gross difference shows its true size. CountLab picks out mix-up pairs in count results, uses the zone map to show where errors cluster, and gives these groups more attention in the cycle count plan.