Abhishek S.
Shipping in public. Listening in private.

Abhishek

I lead women’s Indo-Western & Premium at Max Fashion. I also wrote the AI that runs the buying floor.

Rare profile. Category operator who ships production code.

Senior Buying Leader · Max Fashion Women’s Indo-Western & Premium · 530+ India stores NIFT ’12 · On the women's wear floor

abhishek@bengaluru ~ %
>role: senior buying lead
>dept: women’s indo-western + premium
>floor: 530+ stores india

The Inditex Playbook

Zara’s scarce product is not clothing. It is time left uncommitted. Since its first A Coruña store opened in 1975, the Inditex model has treated a late decision backed by store evidence as more valuable than an early decision backed by a forecast.

How the loop works

Ferdows, Lewis, and Machuca reported in 2004 that Zara could move some designs from sketch to store in about 15 days. Stores received shipments twice a week. The playbook is that loop, not the delivery calendar.

Inditex keeps design, commercial decisions, fabric access, distribution, and store feedback under close control. It does not make everything nearby: predictable basics can travel from lower-cost factories, while uncertain fashion items justify shorter production routes through Spain, Portugal, Morocco, and Turkey. Arteixo acts as the loop’s physical clock.

Conventional seasonal bet Inditex response loop
Commit most volume before launch Reserve capacity for in-season decisions
Optimise unit cost Optimise total exposure
Replenish from forecasts Replenish from store evidence
Accept markdowns as calendar events Treat markdowns as forecast errors

The governing equation comes from concept littles law:

inventory = weekly throughput × flow time

Assumption: at 10,000 units a week, a 12-week flow exposes 120,000 units; a 2-week flow exposes 20,000. The unit cost may rise, but the wager shrinks by 100,000 units.

What the model actually buys

Small batches buy information. A weak style can disappear without a warehouse full of evidence; a strong style can earn a repeat order. Caro and Gallien documented how Zara later added a formal allocation model, tested during 2006, to distribute scarce stock across stores. Software did not replace the commercial loop. It decided where the next box should go.

This is also why copying weekly drops rarely works. Delivery cadence sits downstream of fabric availability, production capacity, allocation rules, and store-level observation. Copy the calendar without those dependencies and the result is merely more frequent lateness.

What’s contested

Researchers broadly agree that quick response reduces forecast exposure. The disputed part is attribution: Inditex does not publish the counterfactual showing how the same assortment would perform under a conventional calendar.

The environmental ledger is also unsettled. Short runs can reduce unsold inventory, yet faster product turnover can increase total production and consumption. Inventory efficiency and material restraint are different claims.

Replication presents a third question. The Iberian production base, Arteixo distribution network, property choices, and decades of operating habit arrived together. A retailer can copy one component in a budget cycle; it cannot purchase the history that made the components fit.

Why this has to do with other realms

Inditex resembles concept bayesian updating expressed through fabric. Each store delivery is a prior; sales, requests, and returns form the evidence; the repeat order becomes the posterior decision. It also resembles concept real options: unused capacity has a cost, but it preserves the right to decide after uncertainty falls.

The open question

If demand prediction becomes cheap but fabric still requires twelve weeks, does advantage move to the better model or remain with the shorter physical clock?

Key Sources

Further Reading

See Also

Abhishek's take

The Inditex playbook is usually told as a story about speed. I read it as a discipline of keeping the purchase order blank until the evidence improves.

Where I’ve used this

On the buying floor, I use models to rank demand signals after sales begin. They can improve the next decision, but they cannot rescue a fabric route that made every decision three months ago. If the model thinks in minutes and the cloth moves in months, which clock is running the business?

Tags: #inditex #zara #fast-fashion #vertical-integration #store-cluster #qr #drop-cadence