ReferenceUpdated 2026-07-06

colour, size, gender and age_group

The Merchant Center apparel attributes, color, size, gender and age_group, accepted values, when they're required, and why consistent values drive filters and variant grouping.

Plain-English summary

For clothing and accessories, these four attributes are how shoppers filter and how Google understands your variants. Google requires them in apparel categories, and it expects consistent, recognisable values, 'navy' and 'Navy Blue' and 'dark blue' scattered across your catalogue quietly cost you filtered impressions. Getting the values normalised is what makes size and colour filters actually work for you.

The apparel attributes

Attribute What it is Example
color The product's colour (base colour, optionally with a name) Navy
size The size, in a value shoppers recognise UK 10
gender Who the product is for female
age_group The age bracket the product is designed for adult

In apparel these four are not optional decoration: they're how Google builds the size and colour filters shoppers use, and how it makes sense of a variant group. A dress family with clean color and size on every variant is filterable and shoppable; the same family with missing or inconsistent values is neither.

Accepted values

  • color: free text, but should lead with a recognisable base colour. Up to three colours can be combined for multi-colour products, separated by / (e.g. Red/White/Blue), with the dominant colour first.
  • size: free text, but consistent within a sizing system (S/M/L, or UK 8/UK 10, or EU 38). Pair with size_type (regular, petite, plus, tall, big, maternity) and size_system (UK, EU, US…) where relevant.
  • gender: one of male, female, unisex. Fixed set.
  • age_group: one of newborn, infant, toddler, kids, adult. Fixed set.

gender and age_group are closed lists: use the exact tokens. color and size are open, which is exactly where consistency problems creep in.

Deep dive Why consistency matters more than any single value

The hidden cost in these attributes isn't wrong values: it's inconsistent values across a catalogue, because it fails silently:

  • Filter drop-out. Shoppers narrow Shopping results by colour and size. If one product says Navy, another Navy Blue, a third dark blue and a fourth #1a2b4c, Google can't reliably bucket them into one "Blue" filter: so products fall out of colour-filtered results without any disapproval to warn you. The same happens with S vs Small vs 8 for size. Every inconsistent value is a lost filtered impression.
  • Variant grouping breaks down. The whole point of a variant group is that the shopper picks colour and size on the page. If the variants' color and size values are missing or garbled, Google can't present a clean picker: see item_group_id and variants.
  • Option-name drift upstream. Store platforms let each product name its own options: "Colour", "Color", "Shade", "Finish" for the same concept. Left as-is, half the catalogue ships without a usable color attribute at all. The values have to be normalised in the feed layer to a single consistent scheme, regardless of what the source option was called.
  • Gender/age mismatches. A kids product mistakenly tagged adult, or gender: male on a women's line, misplaces the product in Google's understanding and can surface it to the wrong audience or the wrong category requirements.

Consistency is worth more than perfection on any single item: a catalogue where every blue thing says exactly Blue outperforms one with beautiful but unique colour names on every product.

How this works in /tools

Enrichment normalises color and size to a single consistent scheme across the whole catalogue (pulling the base colour out of marketing names, unifying option-name drift ("Colour"/"Shade"/"Finish" → color), and standardising size formats) and validates gender and age_group against their accepted values, all written to a supplemental layer that never overwrites your source data. Attribute completeness and consistency are direct inputs to your Feed Health Score. For the normalisation workflow, see colour and size normalisation in Optimise; for how these attributes distinguish variants in a group, see item_group_id and variants.

Frequently asked questions

Are these attributes always required?

They're required in apparel and accessories categories, and the requirement is triggered by the google_product_category you assign. Outside apparel they're optional, but colour and size are still worth sending wherever a shopper might filter by them.

My colours are things like "Midnight" and "Ocean", is that allowed?

The value can carry your marketing name, but it should lead with a recognisable base colour so Google's filters can place it, e.g. "Blue" or "Navy / Ocean". A pure marketing name with no standard colour word can drop out of colour-filtered results.

What values does gender take?

A fixed set, male, female or unisex. age_group is also fixed, newborn, infant, toddler, kids or adult. These aren't free text; use the exact accepted values.

Put this into practice. /tools rebuilds messy product data into Merchant Center-ready feeds. Connect a store and see your Feed Health Score in minutes.
Try /tools →