Article of the weekBusiness Trends

Shein Ships 10,000 New Designs a Day. Zara Took a Year to Copy One.

13 July 20264 min read

Shein didn't beat Zara at fast fashion. It rebuilt the entire supply chain around 100-piece test runs — and turned every customer click into a manufacturing signal.

Zara invented fast fashion. Two-week design-to-shelf cycles. 12,000 SKUs a year. Every MBA case study calls it the gold standard. Shein now ships 10,000 new designs — a day. And it does it with roughly the same headcount Zara had in 2005.

The trick isn't more designers. It's a supply chain rebuilt for one thing: micro-batch testing. Shein's factories in Guangzhou accept orders as small as 100 units. A design goes live on the app, gets impressions, gets clicks, gets add-to-carts. If the signal is strong, the factory produces another 500. Weak signal, the SKU dies in 48 hours. There is no inventory risk, because there is barely any inventory.

Zara can't copy this. Zara's factories in Spain and Portugal run 5,000-unit minimums because they're built for scale, not speed of failure. Their supply chain is optimised for a 60% hit rate on 12,000 SKUs. Shein is optimised for a 2% hit rate on 1.5M SKUs. When only 2% need to work, you can be far more experimental.

The result: Shein's GMV crossed $50B in 2025, with net margins that Zara's parent Inditex can't match on any category. And the moat isn't the app or the price. It's the 5,000+ Guangzhou factories wired directly into Shein's inventory system — a physical network that took 12 years to build and can't be replicated by throwing capital at it.

The lesson for founders: fast fashion wasn't the innovation. Shein rebuilt what 'fast' meant by shrinking the unit of experimentation. In your business, what's the smallest test you could run this week — and what's stopping you from running 100 of them?

From the archive

  • Business Trends · 5 min read

    Nvidia's $10 Trillion Run Was an Accident. Jensen Just Didn't Flinch.

    For 20 years Nvidia sold graphics cards to gamers. Then researchers noticed CUDA could train neural nets — and Jensen bet the company on a market that didn't exist yet.

  • Startups · 4 min read

    Adobe Paid Figma $1B to Walk Away. It Was the Best Deal Figma Ever Did.

    When regulators killed the $20B acquisition, everyone assumed Figma had lost. Two years later it's IPO'd at $35B — and Adobe is still trying to catch up.

  • Comeback Stories · 5 min read

    Boeing Bought McDonnell Douglas in 1997. McDonnell Douglas Won.

    It looked like the biggest aerospace acquisition in history. It was actually a reverse takeover — and it's why 737 MAXes fell out of the sky 20 years later.

  • Business Trends · 5 min read

    AI Coding Agents Are Quietly Shipping Real Startups in 2026

    Solo founders are launching $1M ARR SaaS with no engineers. The new playbook isn't 'AI-assisted' — it's AI-built, founder-edited.

  • Comeback Stories · 5 min read

    LEGO Was 6 Weeks From Bankruptcy. Then It Stopped Inventing.

    In 2003, LEGO was losing $1M a day. The CEO's fix sounded crazy: stop making new products, fire the visionaries, and rebuild the brick.

  • Startups · 4 min read

    Zomato Bought Blinkit for $568M. Three Years Later It's Worth More Than Zomato.

    Everyone thought Deepinder Goyal had overpaid for a dying quick-commerce startup. He'd actually bought India's next billion-dollar category.

Come back next Monday.

A new article drops every week — startups, trends, and comeback stories worth your 5 minutes.

Browse case studies →