The encyclopedia · Strategy & Leadership · Operational decision · 2007–2010
Zara computed every store's replenishment by size and lifted sales 3–4%
An OR model of store size assortments set Zara's shipments from two warehouses to 1,500 stores, adding ~3–4% sales.
Zara (Inditex)
The solution
Zara replenishes about 1,500 stores worldwide from just two central warehouses, and fast fashion's short lifecycles meant stores frequently had the wrong sizes on the shelf. Stockout and markdown losses followed, but store managers had no way to see the demand link for each size.
Zara deployed a new process that uses sophisticated operations research models to determine every inventory shipment it sends to each store. The model takes a size-assortment view of store inventory, connects stock levels to demand, and decides replenishment quantities per store and size, overcoming significant technical and human resistance during deployment.
A controlled field experiment estimated the new process increased sales by 3–4 percent, worth roughly $233 million in profits and $353 million in additional revenues for 2007 and 2008. The work made Caro, Gallien and Zara colleagues Franz Edelman laureates in the 2009 competition.
Why it worked
- The size-assortment view matches how fashion demand actually behaves
- A field experiment proved the lift instead of asserting it
- Centrally computed shipments freed stores from guesswork
What can be applied
In fashion the unit of a stocking decision is the size assortment, not the item; model the link between what is on the shelf and what sells, then test it in the field.
Aftermath
The 2010 Interfaces paper won the Franz Edelman Award competition. Caro's business-school case 'Zara: Staying Fast and Fresh' became widely taught and won the Case Centre's New Case Writer and POM awards.
Sources
- Zara Uses Operations Research to Reengineer Its Global Distribution Process
- Felipe Caro profile
- Zara Uses Operations Research to Reengineer Its Global Distribution Process (abstract)
spotted an error? The archive wants to know.