What item_brand, item_category and item_variant record
These three fields sit alongside the product identifier and name inside a GA4 ecommerce event, and they exist to group products rather than to identify them. Brand records who makes the item. Category records where it sits in your shop’s structure, and GA4 accepts several levels of it, so a deep catalogue can send a top-level section and its subsections rather than flattening everything into one label. Variant records which version of the product this is: the size, the colour, the capacity, the flavour.
All three are optional. GA4 will accept an event without them and the order value will still be right. What you lose is the ability to ask any question broader than one product at a time.
Why item_brand, item_category and item_variant matter
They turn a long list of products into something a business owner can act on. Without them the product report is a flat table sorted by revenue; with them you can see which section of the shop is growing, which brand people browse but rarely buy, and which sizes or colours are worth restocking. For a shop with a wide catalogue that difference decides whether the reports get opened at all.
Variant is the one most often left out and the one that pays back fastest. Returns, stock-outs and abandonment usually concentrate in specific versions of a product rather than the product as a whole, and without the field they are invisible — the parent product just looks moderately disappointing.
Where item_brand, item_category and item_variant go wrong
Drift is the usual fault. The category on the product page comes from the page template, the one in the basket comes from a different database field, and the two disagree — so the same order counts under two sections. The fix is to take all three attributes from one source in the shop database and pass them through unchanged at every step.
Free text is the second problem. Brands entered by hand end up with variant spellings and stray capitalisation, and each becomes its own row in the report. Variant has the same weakness when the colour comes from a supplier’s description rather than a controlled list.
The third is treating category as a marketing label. Filling it with a campaign name or a promotional phrase makes it useless for structural reporting, and it is not the place for it. And if your product genuinely has no versions, leave variant out rather than sending something like “default”, which creates a row that means nothing.
How to act on it
Populate all three from the shop database, using the same values that drive your navigation and your product feed. Consistency between the site, the analytics and the feed is worth more than richness, because it is what lets you compare a GA4 report with what the ad platforms see.
Then read them at the level you can act on. Category performance tells you where to spend attention and where to change navigation, brand performance tells you what to stock and what to negotiate on, and variant performance tells you what to reorder. Once these fields are clean, the same structure carries into your product feed and into ad platform reporting, which is why they are worth agreeing during a GA4 ecommerce setup rather than added later.
