Deere's C&CE cut $890M of inventory and lifted on-time shipments to 92%.

Deere & Company's Commercial and Consumer Equipment division had about $3 billion of sales and 2,500 dealers, but plants shipped late and inventory sat high while service risked falling.

The common fix is to trim inventory everywhere. Deere instead used multistage inventory planning and optimization, a state-of-the-art stochastic inventory approach, to set trustworthy weekly inventory targets by product family.

Those targets were then paired with dealer incentives to convert the division from pushing product to pulling it, so inventory moved only when a real order called for it.

The change lifted factory on-time shipments from 63% to 92% while keeping customer service at 90%, and between 2001 and 2003 Deere reduced or avoided $890 million of inventory, improving annual shareholder value added by $107 million.

Set trustworthy targets, then turn push into pull

Push systems build inventory against forecast error, while targets make pull workable

Multiechelon optimisation handles 100 product families and a 26-week horizon

Trustworthy weekly targets let dealers order what they will actually sell

Dealer incentives tied the new targets to behaviour, not just numbers

Averages hide the problem; the lever is trustworthy per-item inventory targets that let a network run pull, not a single global number.

Deere's factory on-time shipments rose from 63% to 92% with customer service held at 90%. Between 2001 and 2003 it reduced or avoided $890 million of inventory and improved annual shareholder value added by $107 million, and by end-2004 was set to exceed $1 billion a year ahead of plan.