Operations Management Case: Zara's Supply Chain

Case SummaryZara belongs to Spain's Inditex group and is the group's oldest, largest and most successful brand. In 2018 Zara accounted for 69% of group sales (about €18 billion). Its core is a responsive supply chain that can design, produce and put new items on shelves within three weeks. Customers come back often because there are always new styles in store.

In 2010 Zara launched Zara.com. By 2018 online sales were 12% of group sales, reaching 14% in countries with e-commerce operations. But online orders brought high return rates and complex reverse logistics.

Zara's response: - Launched "store optimization": closing small stores and concentrating resources on large flagships. Example: four small stores in Bilbao merged into one flagship, with more space and higher sales. - Rolled out RFID (radio-frequency identification tags) across the board, lifting inventory accuracy from below 70% to above 98%. - Began trying to ship from store inventory instead of relying entirely on distribution centres (DCs). - Launched Click & Collect (order online, pick up in store), which caused in-store congestion; Zara experimented with a robotic sorting system to ease it. - Invested heavily (€1.8 billion in 2017, €1.4 billion in 2018) to become "fully integrated, fully digital, fully sustainable."

Core questions: with e-commerce growing fast, what is the future role of the store? How should inventory be integrated? Should Zara replicate its UK "integrated store + DC" model in other markets?

Class Notes

1. Low-Hanging Fruit: the basics, improved

In the early stages of a supply chain, improvements tend to focus on links that are easy to see and easy to measure:

Better inventory accuracy (Zara used RFID to raise the industry's 70% accuracy to 98%).

More efficient transport and warehousing (for example, integrating DCs and stores to cut duplicate inventory).

Store optimization: merging small stores into big flagships, reducing redundant inventory and improving the experience.

These are the "low-hanging fruit" — the first step in raising efficiency.

2. The real challenge: managing uncertainty

On the slides, the professor particularly stressed: We have to learn to manage uncertainty. In fashion, uncertainty shows up as:

Large demand-forecast errors: forecast error for functional products is around 10%, but for innovative fashion items it can reach 40%–100% (see the table on slide 13).

Short life cycles: 3 months to 1 year for new fast-fashion items, far shorter than the two years for functional products.

High return rates: the average online return rate is 30%, three times that of offline.

Zara's approach is delayed commitment: test with small batches first, then go to large-scale production once market signals are clearer. Shein pushes this logic to the extreme, making 100–200 pieces of each style first before deciding whether to scale up.

3. Responsiveness is harder to copy than efficiency

Zara's core isn't "better equipment than everyone else," but an organization that can respond fast. As a former general manager of Inditex put it:

"We don't understand why others don't do this. Everyone has the equipment, and anyone can copy the processes."

Behind that line lie the barriers of organizational habits, process coordination and culture. Rapid response is very hard to copy wholesale.

Lessons for Arcus42Where is our "supply chain"?

As a multi-family platform, our "goods" aren't clothes, but time, knowledge, vendors and experience.

For Arcus42:

Knowledge = inventory: it needs unique IDs, visibility and reuse (like Zara's RFID).

Vendors = partner factories: they need profiles, a matching engine and a feedback system.

Experience = test orders: pilot at small scale first, then replicate to more family clients.

Client managers = frontline stores: they need to travel light, with the middle office backing them all the way.

In other words, our middle office is Arcus42's "Zara warehouse + system."

AI's new opportunity: from "fast fashion" to "real-time intelligence"

The slides ended with a question: could there one day be a model even faster than "ultra-fast fashion"?

The answer may be AI-driven real-time feedback. For Arcus42:

AI can do "case search" and "solution recommendation" — the equivalent of real-time restocking.

Client behaviour data (questions, needs, document edits) are trend signals.

Through AI routing, the middle office can quickly dispatch work to the best-suited expert or vendor.

This is our own "real-time supply chain."

I always remind myself that doing things well takes enough time, patience and persistence. Operations is work that grinds down your temperament, but I believe there may also be people who are simply born loving these daily routines. Find people like that, and treat them with great enthusiasm and recognition.

Tell them they are important.