How do you merge products from multiple clothing manufacturers into one brand catalog?

NoLabel sources from more than ten blank-apparel manufacturers, including Stanley/Stella and Nameson, each shipping data in its own format. One deterministic import preset reads it all the same way, consolidating 2,305 products and 14,281 variants into a single coherent own-brand line.

How much manual work is importing a spreadsheet with hundreds of columns?

NoLabel's source was a hand-maintained Google Sheet, 265 columns wide with three duplicate headers, reconciled by hand season after season. A deterministic import preset now reads that sheet identically every time; it worked well enough to become the template every other brand preset starts from.

Does an automated content ruleset actually catch bad copy before it publishes?

NoLabel's brand book was codified into machine-checkable rules: banned phrases, unit rules, echo detection and a voice per language. Content quality moved from 43 rejected runs to a clean audit within about five versions of the format document. Nothing reaches the storefront unchecked.

How do you keep product descriptions consistent across thousands of SKUs and several languages?

Every description is generated against the same ruleset, with a per-language voice for English, Dutch and German, and checked before it is applied. That is how all 232 SS26 descriptions went out consistent across 313 collections, instead of drifting per writer and per season.

Do I still have to review every product description by hand?

Not product by product. The ruleset checks every generated description before it is applied, so review effort goes into the rules, not the volume: tune the format document a few versions and the audit comes back clean, as NoLabel's did after about five.

What makes a codified brand book different from a style-guide PDF?

A PDF states intentions; a codified ruleset enforces them. NoLabel's banned phrases, unit rules and echo detection run against every single description automatically, in all three languages. The brand book stopped being advice for writers and became a gate nothing skips.

NoLabel × Automated Commerce

NoLabel is an Amsterdam menswear brand built on publishing the true cost behind every price tag.

Its whole catalog lived in a single Google Sheet, 265 columns wide. Automated Commerce turned that sheet into a repeatable system, and the import preset built for it became the template every other brand in the platform now starts from.

About NoLabel

NoLabel is an Amsterdam menswear brand founded in 2013. It sells minimalist essentials: shirts, knitwear, trousers, outerwear and footwear, styled as one coherent line. NoLabel sells in English, Dutch and German.

The challenge

NoLabel's source of truth was a hand-maintained Google Sheet, 265 columns wide, including three duplicate header columns. The catalog is sourced from more than ten blank-apparel manufacturers, among them Stanley/Stella, Stillshirt, Nameson and Euroralex, each shipping product data in its own format. Yet the storefront has to read as one coherent own-brand line, not a patchwork of ten different suppliers. Reconciling that sheet by hand, season after season, did not scale.

How Automated Commerce helps

Automated Commerce built a deterministic import preset that reads NoLabel's spreadsheet the same way every time, whichever supplier's data sits behind it, duplicate headers and all. That preset became the template every other brand preset in the platform is now copied from. NoLabel's brand book was codified into a machine-checkable content ruleset: banned phrases, unit rules, echo detection and a per-language voice for English, Dutch and German. Every generated description is checked against that ruleset before it is applied, so nothing reaches the storefront unchecked.

The result

The numbers behind the switch:

2,305 products and 14,281 variants run through one deterministic import preset.
More than 10 supplier brands consolidated into a single own-brand line.
313 collections structured across the catalog, sold in English, Dutch and German.
All 232 SS26 product descriptions generated and applied.
Content quality moved from 43 rejected runs to a clean audit within about five versions of the format document.