Product Data for Lingerie & Swimwear: Modelling Cup Sizing and Sets Cleanly

Most apparel has one size axis. A T-shirt is S, M or L; a pair of trousers is a waist number. Lingerie and much of swimwear are the exception — and the exception is expensive to get wrong. Here, a single bra style doesn't have a size, it has a two-dimensional grid: a band (the underbust measurement) crossed with a cup (A, B, C, D and up). Add a bikini sold as a set with independently sized top and bottom, and you have the two data problems that a standard apparel model simply wasn't built for.

Product data for lingerie and swimwear is defined by two complications: a two-dimensional band-plus-cup size logic, and sets that behave as one article with independently sized parts. This is a focused sub-branch of the wider fashion retail challenge — the variant-heavy world of apparel, pushed to its most complex corner.

Why is the size logic for lingerie and swimwear a special case?

The trouble starts with the sheer combinatorics. A normal apparel style has maybe six sizes. A bra style spans a band range (65–100) crossed with a cup range (A–H) — and cup is not absolute: an 80B and a 75C share a cup volume but differ in band. That is why you can't flatten it into a single dropdown:

• Two dependent axes: band and cup vary together, producing 40–50 valid combinations per style — and a matching number of GTIN/EAN keys to track.
• Sister-size overlap: some combinations are near-equivalent, which matters for merchandising and stock but is invisible in a flat size list.
• Swimwear inherits it: supportive swimwear uses the same band/cup logic; simpler pieces revert to S/M/L — so one supplier's range mixes both models in one file.
• Notation chaos: the same size arrives as 75B, 75/B, or split across two columns — and sometimes only inside the article title.

Model this as a flat size field and you either drop valid combinations or list sizes that don't exist. It needs a real two-dimensional variant structure, and getting there means consolidating and normalizing inconsistent supplier data first.

How do you model a bikini or lingerie set as one article?

The second complication is set building. A bikini or a lingerie set is a single sellable article, but its top and bottom can carry different sizes — a 75C top with an M bottom is a normal, common combination. There are two wrong ways to handle this and one right one:

• Two separate products (wrong): breaks the buying experience and doubles the maintenance; the customer has to find and pair two SKUs.
• One flat variant (wrong): collapses top and bottom into a single size, losing the very flexibility that defines a set.
• One parent, two linked size dimensions (right): the set stays one product, but top size and bottom size are kept as separate, connected axes — so every valid pairing exists and impossible ones don't.

The set problem and the band/cup problem are really the same problem: a single article that needs more than one independent size dimension. Solve that once, and both cases fall out of it.

Which standard applies — and where does it stop?

There is no dedicated classification for cup sizing the way FEDAS exists for sport or ETIM for electrical. Lingerie and swimwear ride on the general apparel data stack — and that stack treats size as one field. Here's what the standards and feeds deliver, and where they run out:

So even where a clean feed exists for the branded core, the two hard parts — the two-dimensional sizing and the set structure — are exactly what the standard doesn't carry. That gap is where the manual work lives.

How does Productbay solve this?

The mechanism is linked attributes: instead of one flat size list, Productbay keeps two connected size dimensions on a single article — band and cup, or top and bottom of a set. From there the usual three-step job applies:

• Consolidate: import every source once — supplier CSV, Excel, BMEcat, feed URL, FTP, API — and match by SKU or GTIN/EAN so existing articles update and new ones are created, with the size matrix intact.
• Enrich: AI parses band and cup out of inconsistent notation (even from a PDF datasheet or the article title), normalizes it into a clean two-dimensional structure, writes descriptions, assigns categories and fills gaps from whitelisted sources — always with a review queue before anything publishes.
• Publish: two-way sync to Shopify and Shopware, ERP connections (Xentral, weclapp), and feed exports for Amazon, OTTO and Kaufland — each with per-channel transformations.

The point is that complex variants — band/cup grids and mixed-size sets — stop being a special case that fights your system, and become a modelled structure your catalog understands. Productbay is built for specialist retailers running multi-supplier, multi-channel catalogs, and this sits inside the broader fashion category.

Frequently Asked Questions