Case Study • AI Inventory Markdown Optimization
We caught 47 SKUs before the fire sale
- Client
- National off-price retailer
- Industry
- Off-Price Retail
- Service
- AI Solutions|Data Analytics
Overview
ClearanceIQ is the AI command centre Sphere built for a national off-price retailer. This representative Summer 2026 clearance season follows six stores and shows how the system catches stuck inventory weeks before it becomes a last-week, deepest-discount write-off.
Why does inventory lose margin before teams act?
The longer a unit sits, the more it costs to store and the deeper the eventual markdown has to be. Most teams do not notice a SKU is in trouble until it is already deep into markdown territory.
A SKU stops moving
Historically noticed when someone walks past the rack—often months into the season.
Markdown depth becomes guesswork
A blanket store policy gives up margin on items that never needed the deepest discount.
Six stores make six judgment calls
Regional decisions vary even when the underlying inventory problem is the same.
Our Solution
ClearanceIQ scores inventory continuously, turns every flag into a recommended action, and models the financial effect before a price tag changes.
1. Catch the SKU while there is still room to act
The system caught 112 cold-shoulder tops at day 70, selling at two units a week against a 21-unit peak. It recommended a 40% markdown to $9.99 and a move to the entrance display, projecting $5,280 in recovery versus doing nothing.
2. Price the markdown before approving it
For 84 linen blazers, the Markdown Simulator tested discount depth from 0% to 60% against units sold, revenue, and margin. At the recommended 30% discount, it projected $1,505 in recovered revenue with 49% margin intact.
Key Achievements
Margin protected
Healthy sellers stay at full price while only genuinely stuck items are discounted.
Cash and space recovered faster
Inventory clears in 38 days instead of the 67-day baseline.
Losses taken earlier and smaller
The model acts before the only remaining option is a final-week fire sale.
One policy across the fleet
The same scoring and markdown math runs consistently across every store.
A defensible number before the decision
Finance sees projected recovery and remaining margin before a price changes.
Results
Across the six-store cluster, ClearanceIQ flagged 47 at-risk items and identified $41,200 in projected recovery if the recommended actions were taken that day. Average time to clear a flagged SKU fell to 38 days from a 67-day baseline—nearly halving the time inventory occupied floor and backroom space without earning.
The store cluster, season, and dates are a representative scenario run through Sphere’s anonymized RetailerX Clearance build. Figures are modeled system output, not disclosed production financials.
