Zalando Apparel & Accessories Catalog Data

Each record is one purchasable article — brand × style × colour × size — with its live storefront attributes attached. Three rows from the catalog, exactly as they land in your delivery:

Adidas | 4065428123456 | Shoes > Sneakers > Low-top sneakers | €89.95 | black | sizes: EU 42, EU 43, EU 44 | Upper: synthetic; lining: textile; sole: rubber BOSS | 4062547890123 | Men > Clothing > Shirts > Business shirts | €119.00 | light blue | sizes: S, M, L, XL | 100% cotton Zign | 4059372123456 | Women > Bags > Shoulder bags | €49.95 | cognac | sizes: One size | Outer: imitation leather; lining: textile

Every field above resolves to the dictionary below. No nulls dressed up as zeros, no brand names split across three spellings — one row, one article, one EAN.

The core record carries brand, EAN/GTIN, category tree, price with discount state, colour, size run, declared material composition, aggregate customer rating, and any sustainability badge shown on the product page. Nine structured fields per article is enough to answer most assortment and pricing questions without joining a second source.

Geography: 29 European markets, each served by its own country storefront — Germany, France, Italy, Spain, the UK, the Nordics, Benelux, CEE and more.

Temporal: continuously current catalogue. Pricing history accrues from repeated capture, so a weekly cadence builds a usable discount-depth curve within a quarter.

Granularity: article/variant level — brand × style × colour × size — not brand rollups, not category averages.

API, files, or your warehouse. Daily, weekly, or hourly.

Pick the channel that matches how your team already works: REST endpoints for live lookups, flat files for batch analysis, or a direct pipe into Snowflake, BigQuery or Redshift. Cadence is your call — set it once, change it whenever your models change.

Every delivery ships with the full field dictionary, sample rows for validation, and a schema that does not drift underneath you between refreshes.

Pricing and promotion analysts at apparel brands benchmark their own price points and discount depth against the largest assortment in European fashion e-commerce, market by market. A 30% markdown in Berlin reads differently when you can see whether the same article is at 15% in Amsterdam.

Assortment and buying teams map category-tree breadth against their own range plans: which brands are adding styles in women's outerwear this season, which colour families dominate each price band, where a 7,000-brand platform chooses to go deep versus wide.

Investment and strategy researchers treat the catalogue as a demand-side signal — active style counts, discount pressure and rating distributions across 29 markets, tracked over time rather than guessed once.

Data science teams use the structured attributes as clean training or enrichment input: category classification from category trees, price-elasticity features from price plus rating, sustainability positioning from badge coverage.

Scale first: 62 million active customers, more than 7,000 brands, roughly EUR 17.6 billion in gross merchandise volume across 29 markets. Breadth like that makes the platform a de facto reference price and assortment index for European fashion.

Structure second: product pages follow one consistent template, so brand, EAN, price state, colour, size and material composition arrive as comparable columns instead of free text pulled out of twenty different layouts. Comparability is the hard part of retail benchmarking; here it comes built in.

And a note on what this dataset is not: it is not a historical archive. The catalogue is live commerce, so history exists only where someone captured it — which is precisely why teams start a capture cadence before they need the curve.