GTM. Variable transformations. Normalize messy item data before it hits GA4.

What this means for your store

Merchandising systems, marketplace feeds, and legacy PIM exports rarely agree on SKU format - but GA4 expects consistent item_id values. Variable transformations (variable settings → Format Value) run regex replace, lowercase, substring, and JSON operations on variable output before tags consume them. Clean data in GTM instead of patching every theme template.

Scenario on a real storefront

Product pages push item_id as SKU-12345-US on the US site and sku_12345_us on the UK mirror. You normalize in a Data Layer Variable before GA4 tags fire:

Variable: DLV - item_id (raw)
  Format Value → Transformations:
    1. Regex replace: ^[Ss][Kk][Uu][-_]? → (empty)
    2. Regex replace: -[A-Z]{2}$ → (empty)   // drop -US suffix
    3. Lowercase

GA4 Event tag view_item:
  items.0.item_id = {{DLV - item_id (normalized)}}

// Alternative: Table lookup for legacy SKU → canonical ID map

What to do next

• Apply transformations on variables, not inside every tag - one place to update when merchandising changes SKU rules.
• Document regex rules in the workspace description; opaque transforms confuse the next analyst.
• Use Preview Variables tab on a product URL before wiring purchase items - confirm output matches your ERP.

Bottom line

Transformations keep GA4 item parameters comparable when the catalog is messy. Centralize cleanup in variables so tags stay simple and revenue-by-SKU reports stay trustworthy.