Knowledge — Inventory Problems

Promotion stock planning that protects both ends

A promotion compresses months of demand change into days. Plan it by average and you get empty displays in your best stores mid-wave — and pallets of leftovers everywhere else when it ends.

Why it happens

Initial allocation is a guess

Splitting promo stock by store size ignores who actually sells during promotions. Winners run dry in days; slow stores keep the rest.

Nobody steps ordering back down

After the wave, order quantities stay promotional. The leftover mountain is built in the two weeks after the promotion, not during it.

Promo history distorts normal targets

If promotional spikes flow into regular calculations, the system over-orders for months after a single campaign.

What good control looks like

Stock target through a promotion: raised a replenishment cycle before it starts, returned to standard before it ends — The target goes up one replenishment cycle before the promotion starts, and back to standard before it ends — so the shelf holds through the wave and the tail sells out instead of sitting there.

Allocate a starting split, then follow real sell-through — move stock to where the promotion actually works, during the promotion

Keep promotional demand separate from baseline so regular targets stay honest

Step ordering down automatically the day the promotion ends

Review each campaign: lost sales during, leftovers after — both measured, both improvable

Both ends — no lost sales during the promotion, no overstock mountain after it

Horizon treats a promotion as a temporary layer on top of normal targets: allocation follows sell-through during the wave, the layer disappears on the end date, and baseline targets never get polluted by the spike.

Common questions

Why do promotions always leave overstock?

Because ordering rarely steps down when the promotion ends — quantities stay promotional for weeks. Automatic end-dating of the promo layer removes the leftover mountain.

How do I keep promo spikes from distorting future orders?

Keep promotional demand separate from baseline demand. The spike belongs to the campaign, not to the product's normal target.

See this on your own data.

The free demo begins with a real simulation on your history — where sales, cash and time are leaking, and what the system would have done instead.

Start Your Free Demo