Skip to the data

Bellwether HardwareReorder

As of Thu 6 Aug 2026, 14:20

Signed in Tobias Renn

Sample data

Count these

Which figures is the reorder list trusting that it should not be?

Lines counted

5 of 12

over the last three weeks — the watched lines are counted on a rolling four-week cycle

The other 9,388 SKUs are counted quarterly, by aisle.

Book was wrong on

4 of 5

which is normal, and is the reason for counting rather than a reason for alarm

Wrong enough to change the order

2 of 5

the only figure on this screen that costs anything

Drift that does not move the suggestion is bookkeeping, not a problem.

What gets counted, and when

  1. 1Every watched line, on a four-week rolling cycle, a few at a time on a Tuesday.
  2. 2Any line where two consecutive orders arrived and the shelf still emptied.
  3. 3Any line a colleague reports as visibly wrong, immediately and without a form.
  4. 4Everything else quarterly, by aisle, with the shop open.

Counted figures replace the book figure outright. There is no approval step, because a count that has to be argued for is a count nobody bothers doing.

Which direction is the dangerous one

Book high, shelf low. The list then believes there is stock that is not there, orders nothing, and the line runs out on a Saturday. Book low is the harmless one — it over-orders, which costs money and shelf, and everybody notices immediately.

Two of the five below are book-high, and both of them are lines the list would have got wrong. Bars in both directions are drawn the same weight for that reason: the direction is carried by the side of the centre line and by the word in the row, not by a colour deciding which one a reader should feel worse about.

Counted lines · most drifted first

Each counted line with the book figure, the counted figure, the drift between them, the days of cover the counted figure gives, and whether the drift changed the suggested order.
LineBookCountedDriftCover nowEffect on the orderCounted on
Construction adhesive, 10 ozADH-77144147+6 · book was low15 dNo changeBoth figures suggest 0 cases.28 Jul
#8 × 1¼ in. deck screw, 5 lbFST-0910149-5 · book was high5.7 dOrder changedThe book figure suggested 2 cases; the count makes it 3.28 Jul
Bird seed, black oil sunflower, 20 lbGRD-2264128-4 · book was high4 dOrder changedThe book figure suggested 4 cases; the count makes it 5.21 Jul
½ in. copper couplingPLB-22143034+4 · book was low5.8 dNo changeBoth figures suggest 2 cases.21 Jul
15 A tamper-resistant receptacle, ivoryELE-44712222agrees24.1 dNo changeBoth figures suggest 0 cases.14 Jul

No count on this screen is attributed to whoever did it. The moment a drift has a name attached, the next one gets quietly corrected on the shelf instead of recorded, and the only honest picture of the stock disappears.

All dashboards — back to the dashboard index