The diagnosis for the weekly forecast with promotion lift, the way most systems are set up. We split the promotion-week miss into what the forecast got wrong and what the empty shelf lost, check the supply side, follow the lost units to their cause, then replay the week to prove it.
Step 1. Split the miss
What shoppers wanted on the empty days is estimated from the item's sales rate before the shelf ran out and from comparable promotions. In client data we show this estimate as a range.
Step 2. Check supply, then follow the lost units
- Yes
Was the shelf actually empty? —Data: store on-hand inventory
- Yes
Did deliveries arrive on time and in full? —Data: store receipts against store orders
- No
Was the order big enough for the weekend? —Data: store orders, store POS
- Gap
Why was it too small? —Data: promotion calendar with offer type, sales from past promotions
Step 3. Replay the week with the lift planned by offer type
Sales missed, promotion week: as it happened → lift by offer type with pre-build—
Extra units on hand going into the next week
Same shoppers, same deliveries, same four weeks of sales. Only the promotion plan changes, so the difference belongs to the plan.