id

Almost every operational catalog failure traces back to the id field at some point. The id is what your pixel sends, what your sale_price updates target, what your audit reports key against, what your refund flow looks up. When two systems disagree on what an id means — Shopify variant IDs in the pixel, parent product IDs in the feed, for instance — every other piece of catalog hygiene downstream suffers. The pixel match rate drops below 90%, Advantage+ optimisation degrades, retargeting audiences misfire, and the symptoms look like creative problems or audience problems when the root cause is upstream in the id layer.

The Shopify-specific id problem deserves its own paragraph because it's so common. Shopify ships pixel events with variant IDs as content_ids by default, but many feed-generation tools default to product IDs (the parent SKU). The two formats look similar but never join. The Catalog Match Rate in Meta's Events Manager drops to whatever percentage of your traffic happens to also match by coincidence — typically 40-70%, not the 95%+ that healthy DPA optimisation requires. The fix is to align both ends: either configure the pixel to send product IDs, or change the feed to use variant IDs. WooCommerce and Magento have similar but channel-specific variants of this problem.

For multi-region catalogs, the id needs to handle the same SKU appearing across regions. Two patterns work. Region-scoped feeds (one feed per country with the same id across them) treat the same SKU as one item that ships to multiple destinations; this is cleanest but only works when product attributes are genuinely identical across regions. Region-suffixed ids (`SKU-001-UK`, `SKU-001-US`) treat each as a distinct item, which is the right answer when the same product genuinely has different specs per region — different size labels, different regulatory compliance, different bundles.

Length matters in subtle ways. Google's 50-character limit on id is universally enforced; longer ids get truncated, which silently breaks pixel matching. Stay under 50 with margin. Avoid spaces (channels accept them but the URL-encoding causes downstream issues with pixel pings), avoid case mixing (`SKU-001` and `sku-001` are different items), avoid special characters beyond hyphens and underscores. A 20-30 character alphanumeric-with-hyphens id has all the advantages and none of the failure modes.

The canonical fix is to decide on one id format and enforce it everywhere. That format should be stable (don't include data that changes — sale_price, stock state, anything time-sensitive), unique (no collisions across variants), and consistent across channels (the same SKU has the same id in Google, Meta, TikTok, Pinterest). Most teams that have wrestled this into shape end up using a deterministic format like `{brand-prefix}-{variant-suffix}`, which gives them a visible logical structure plus easy debugging when something goes wrong.

The other often-overlooked rule: once an id is in use, never reuse it. Even after a product is discontinued and removed from the feed, channels retain historical performance data tied to that id for 6-12 months. Reusing a discontinued product's id for a new product starts the new product with the old one's signal — almost always negative because discontinuing usually means it underperformed. The cost of suffixing new SKUs with version numbers is rounding error compared to the cost of inheriting bad signal.