You exported your catalog to every channel and assumed identical data would work everywhere. Now Google Merchant Center is flagging missing GTINs, Meta rejected your variant structure, TikTok Shop cannot parse your categories, and Amazon's listings are stuck in search limbo. One catalog does not mean one feed.
What you'll take away
- Different schemas — Google, Meta, TikTok and Amazon each define required and optional attributes differently, so identical exports break.
- Channel-specific rules — A feed-management system applies transformations like title restructuring and category mapping per destination.
- Variant handling — Each platform structures sizes, colours and parent-child relationships differently, and mismatches cause silent failures.
- Source of truth — Treat your ecommerce catalog as canonical data, then layer rules on top instead of maintaining four manual exports.
Why can't I just export one product feed to every platform?
Each advertising and selling platform defines its own product data schema, category taxonomy and business rules, so an identical feed export will fail validation or perform poorly on at least one channel. Google Merchant Center requires the Google product category attribute and a valid GTIN for most items. Meta catalogs use Facebook's own category structure and content type fields. TikTok Shop has its own catalog schema and category tree. Amazon's listing API works with entirely different inventory and variation systems.
None of these schemas are interchangeable. A Google product category code like "Apparel & Accessories > Clothing" does not map to Meta's category tree without transformation. A GTIN that satisfies Google's identifier requirement might not even be a field Amazon expects in the same position.
From the trenches — The most common breakage we see is a merchant copying their Google Shopping feed URL into Meta Catalog Manager, then watching 40 percent of items fail because the content_type field is missing and the Google category does not map to a valid Facebook product category.
What does each platform actually require?
Google, Meta, TikTok and Amazon each demand different attribute sets, category formats and identifier rules. Google Merchant Center expects a structured XML or CSV feed with attributes like google_product_category, gtin, mpn and availability, and Google's product data specification allows 150 characters for the title field. Meta catalogs require content_type and use their own product category enum. TikTok Shop expects category IDs from its own taxonomy and has specific image requirements. Amazon uses its own product type and browse node system entirely outside any feed schema.
| Platform | Category system | Key required attributes | Title limit |
|---|---|---|---|
| Google Merchant Center | Google product taxonomy (numeric) | google_product_category, gtin, brand, availability | 150 characters |
| Meta Catalog | Facebook category enum | content_type, retailer_id, availability, condition | 200 characters (recommended) |
| TikTok Shop | TikTok category IDs | category_id, product_id, inventory, status | 255 characters |
| Amazon | Browse nodes and product types | product_type, sku, asin (if existing), inventory | 200 characters (indexable) |
How do I handle category mapping across channels?
Category mapping is the process of translating your internal product taxonomy into the specific category format each platform expects, because no two channels share a category system. Your store might have a category called "Men's T-Shirts." Google wants a numeric path like "2124" (Apparel & Accessories > Clothing > Shirts & Tops). Meta wants a text enum like "apparel_and_accessories". TikTok wants its own category ID from a completely separate tree. Amazon maps via browse node IDs tied to product types.
Mapping is not guessing
Each platform publishes its own taxonomy file. Google's Google product category taxonomy has roughly 6,000 nodes. Meta's catalog category enum has a few hundred. A feed-management system stores these taxonomies and applies a channel-specific category mapping rule that translates your internal category to the correct ID or path per destination, so you set the mapping once and it applies on every sync.
What happens to product variants on each platform?
Each platform structures variants differently, and a mismatch between your variant data and the channel's expected format causes items to list as separate products or fail entirely. Google Merchant Center uses item_group_id to tie variant children together. Meta catalogs use retailer_id on the parent and retailer_product_group_id on children. TikTok Shop has its own variant SKU structure. Amazon uses parent-child relationships via product type definitions.
If your variant handling exports a parent product with child variants but the channel expects flat individual items, you get orphaned listings. If you send child items without a group identifier, Google shows them as unrelated products instead of a size or colour family.
Before · Google feed with broken variants
title: Blue T-Shirt M\ngtin: 5060123456789\nitem_group_id: (empty)
After · Variants grouped correctly
title: Gildan Men's Cotton T-Shirt, Navy, Medium\ngtin: 5060123456789\nitem_group_id: TSHIRT-NAV-001\nsize: M\ncolor: Navy
Should I restructure product titles per channel?
Yes. Title structure should be channel-specific because each platform's search algorithm and display format prioritise different keyword placements. Google Shopping truncates titles at 150 characters and its algorithm weights the first 70 characters heavily for query matching. Meta's dynamic ads pull titles into compact card formats where shorter titles perform better. Amazon indexes up to 200 characters but weights the first 50 for relevance scoring. TikTok Shop displays titles in a narrow mobile card where brevity matters.
A channel-specific title rule can restructure the same source title into different formats per destination. For Google, you might lead with brand, then colour, then product type, then size. For Amazon, you might lead with the most searched keyword phrase. For Meta, you might strip attributes and lead with the product name. The source data stays the same. The transformation rule changes per channel.
Before · Same title sent everywhere (29 characters)
Blue T-Shirt M
After · Google (76 characters)
Gildan Men's Short Sleeve T-Shirt, Cotton, Navy, Medium
After · Meta (44 characters)
Gildan Navy T-Shirt, Medium
Title rules are not optional polish
Google Shopping titles that exceed 150 characters get truncated in auction results. Titles under 30 characters miss keyword coverage. A title transformation rule that restructures source data per channel is the single highest-impact feed optimisation after category mapping.
How does a feed-management system solve this?
A feed-management system treats your ecommerce catalog as the source of truth and applies channel-specific rules on top, so you maintain one canonical dataset and transform it per destination rather than maintaining four manual exports. The system reads your catalog, applies rules for category mapping, title restructuring, attribute filtering, variant grouping and exclusions, then outputs a feed that matches each platform's schema.
What does it look like in practice — a worked example
Say you sell a cotton navy t-shirt in sizes S, M, L and XL. Your store catalog has one parent product with four variant children. Here is how that single product gets transformed for each channel.
| Attribute | Google Merchant Center | Meta Catalog | TikTok Shop | Amazon |
|---|---|---|---|---|
| Title format | Brand + product type + material + colour + size | Brand + colour + product type + size | Product type + colour + size | Keyword phrase + brand + colour + size |
| Category | 2124 (numeric path) | apparel_and_accessories | TikTok category ID | Browse node ID |
| Variant grouping | item_group_id | retailer_product_group_id | Variant SKU structure | Parent-child product type |
| Identifier | gtin required | retailer_id required | product_id required | sku or asin |
| Availability | in_stock | in stock | inventory count | inventory quantity |
That is one parent product producing four different feed structures. If you export one CSV to all four, at least two will have wrong category values, at least one will have broken variant grouping, and Amazon will not accept the file format at all.
Common questions
Can I use the same product feed for Google Shopping and Meta ads?
You can technically send the same CSV or XML file to both, but it will underperform or break. Google requires its own numeric product category and a valid GTIN for most products. Meta uses a different category enum and a content_type field Google does not use. A feed that validates in Google Merchant Center will likely fail Meta's catalog review or lose category targeting, so you need channel-specific transformations.
What happens if my product category does not exist in Google's taxonomy?
If your exact product category does not exist in Google's product category taxonomy, you map to the closest parent category that does exist. Google validates the google_product_category field against its own taxonomy file, so a value outside that file causes a disapproval. A feed-management system lets you define a fallback mapping rule, so items without an exact match roll up to the nearest valid parent category automatically.
Do I need different feeds for TikTok Shop and Amazon?
Yes. TikTok Shop uses its own catalog schema with category IDs from its internal taxonomy and a variant structure based on SKU-level inventory. Amazon uses product type definitions and browse node IDs from its own system, and it does not accept standard feed formats like XML or CSV in the same way. You need channel-specific outputs for both, even if your source catalog is identical.
How often should I sync my feed to each channel?
Google Merchant Center fetches your feed on a schedule you set, typically daily. Meta catalogs can sync via scheduled fetch or real-time API updates. TikTok Shop and Amazon each have their own refresh and API rate limits. A feed-management system handles sync timing per channel automatically, so you are not managing four different schedules manually.
What is the biggest risk of sending the same feed everywhere?
The biggest risk is silent failure. Your items may be accepted but mapped to the wrong category, have broken variant grouping, or carry titles that truncate or underperform in search. You will not see errors in the dashboard. You will see it in wasted ad spend, low impression share, and poor feed quality scores that limit auction performance across every channel simultaneously.