Google Merchant Center CSV for AI Shopping Readiness
> A Google Merchant Center CSV feed is a structured product data file formatted to Google's product data specification — exported by FeedBridge as a live feed format — that submits product identity, pricing, availability, and attribute data to Google's merchant infrastructure, which powers both Google Shopping and Google AI Mode product surfaces.
---
What Is a Google Merchant Center CSV Feed?
A Google Merchant Center CSV feed is a tabular data file containing product information formatted to Google's product data specification. Each row in the file represents one product; each column represents a product attribute. The file is submitted to Google Merchant Center — Google's product data platform — where it is validated, indexed, and used to power product listings across Google's surfaces.
Google Merchant Center is the central product data system for all Google Shopping activity. It ingests product feeds, validates them against its attribute requirements, and distributes approved product data to Google Shopping, Google Search product listings, YouTube Shopping, and — increasingly — Google's AI-powered shopping surfaces including Google AI Mode. For merchants, Merchant Center is the required gateway: no Merchant Center feed means no presence on any Google Shopping surface.
The CSV format (comma-separated values) is one of the accepted submission methods alongside XML and Google Sheets. It is the format most commonly used in automated feed pipelines because it is compact, portable, and straightforward to generate programmatically. FeedBridge exports Google Merchant Center CSV as a live feed format, generated from the same normalised product catalog that also powers the ACP JSON-LD feed, UCP Interactive Protocol, Meta Commerce Manager CSV, and Amazon Inventory File TSV outputs.
Understanding what goes into a well-constructed Merchant Center CSV — and what the gaps cost — matters more as Google's AI-powered surfaces become the primary discovery layer for product search.
---
Required Fields in the Google Merchant Center CSV
Google's product data specification defines a set of required attributes that every product record must include to be approved in Merchant Center. Products missing required fields are disapproved and will not appear on any Google surface. [web:35][web:39]
| Attribute | GMC Field Name | Notes | |---|---|---| | Unique product identifier | `id` | Merchant-assigned; must be stable across updates | | Product name | `title` | Max 150 characters; must match landing page | | Product description | `description` | Max 5,000 characters; plain text or basic formatting | | Product page URL | `link` | Direct URL to the product's landing page | | Primary image | `image_link` | URL of the main product image | | Price | `price` | Numeric with currency code (e.g., `29.99 USD`) | | Availability | `availability` | `in_stock`, `out_of_stock`, or `preorder` | | Condition | `condition` | `new`, `refurbished`, or `used` | | Brand | `brand` | Required for all new products except movies, books, and musical recordings | | GTIN | `gtin` | Required for branded products with an assigned GTIN |
These ten fields form the minimum viable product record for Merchant Center approval. [web:39][web:35] Every field has format requirements: `price` must include the currency code; `availability` must use one of the three accepted values exactly; `gtin` must be a valid GS1-registered barcode. Formatting errors — such as a price submitted as "₹2,999" instead of "2999 INR" — cause product-level disapprovals.
---
Strongly Recommended Fields That Affect AI Surfaces
Beyond the required fields, Google's product data specification includes a set of attributes that are optional in terms of feed approval but practically essential for performance on AI surfaces and Shopping — particularly for variant products and category-specific listings. [web:35][web:43]
| Attribute | GMC Field Name | Why It Matters | |---|---|---| | Manufacturer Part Number | `mpn` | Required when GTIN is not available; completes unique product identity | | Item group ID | `item_group_id` | Links product variants; required for variant products in major markets including the US, UK, Germany, France, Japan, and Brazil | | Colour | `color` | Variant differentiator; required in item groups with colour variants | | Size | `size` | Variant differentiator; required in item groups with size variants | | Product category | `google_product_category` | Google's taxonomy path; improves query matching | | Sale price | `sale_price` | Communicates promotional pricing; requires `sale_price_effective_date` | | Sale price effective date | `sale_price_effective_date` | Bounds the promotional period; without this, sale price may not be shown | | Additional image links | `additional_image_link` | Supports multi-image product listings | | Product type | `product_type` | Merchant-defined category path; supplements `google_product_category` |
For AI Mode in particular, the richness of the product record determines how well the AI surface can match the product to natural language queries, filter by specific attributes, and surface relevant products to buyers with nuanced intents. A product with only the minimum required fields will be approved, but a product with fully populated optional attributes — complete variant structure, sale pricing with valid date bounds, category taxonomy, and multi-image support — gives the AI surface significantly more signal to work with.
---
How AI Mode Uses Google Merchant Center Data
Google AI Mode is a UCP-enabled surface that uses product data from Merchant Center (among other sources) to answer product-related queries, recommend specific items, and surface transactable products to buyers. The Merchant Center CSV feed contributes to AI Mode product data in two ways.
First, it is the source of record for the core product attributes — identity, pricing, availability, and structured variant data — that AI Mode uses to evaluate and display products. A product that is correctly submitted to Merchant Center with complete attributes is a product that AI Mode can confidently display, price, and link to purchase.
Second, Merchant Center data interoperates with UCP-based product data for merchants who implement both. For merchants on FeedBridge, the UCP Interactive Protocol endpoints expose live product data to Google AI Mode via real-time REST queries, while the Merchant Center CSV provides the structured product catalogue index that underpins Merchant Center's product matching and approval system. The two are complementary: the Merchant Center CSV gives Google the complete product catalogue index; the UCP endpoints give AI Mode real-time access to current availability, pricing, and cart operations.
This means that AI shopping readiness for Google surfaces has two data layers: the Merchant Center CSV (the catalogue index, processed in batch) and the UCP Interactive Protocol (the live query layer, processed in real time). A well-constructed Merchant Center CSV is the foundation — it determines whether the product is approved and indexed. UCP interactive support is the live layer — it determines whether AI Mode can query the product in real time and handle cart sessions.
---
FeedBridge CSV Export and Field Mapping
FeedBridge exports Google Merchant Center CSV as a live feed format. The export is generated from the normalised product record in FeedBridge's catalog, mapping each field in the enriched product record to the corresponding GMC attribute. FeedBridge's auto-field mapping feature handles 13 common columns — including SKU, title, price, GTIN, MPN, and images — as part of the catalog import workflow, meaning the source data for the GMC CSV export is aligned to GMC field requirements from the point of catalog ingestion.
The following mapping illustrates how FeedBridge's internal product fields correspond to GMC attribute names:
| FeedBridge Field | GMC Attribute | Notes | |---|---|---| | `item_id` | `id` | Direct mapping | | `title` (AI-enhanced) | `title` | FeedBridge AI-enhanced title mapped to GMC title field | | `description` (AI-enhanced) | `description` | AI-enriched description maps to GMC description | | `product_url` | `link` | Direct mapping | | `image_url` | `image_link` | Primary image maps to GMC primary image | | `price` | `price` | Numeric value with ISO 4217 currency | | `currency` | Combined with `price` | Currency appended to price value per GMC format | | `availability` | `availability` | Direct mapping (`in_stock` / `out_of_stock`) | | `brand` | `brand` | Direct mapping | | `gtin` | `gtin` | Direct mapping | | `mpn` | `mpn` | Direct mapping | | `sale_price` | `sale_price` | Maps to GMC sale price field | | `sale_price_start_date` / `sale_price_end_date` | `sale_price_effective_date` | Combined per GMC format requirements | | `category` (normalised) | `google_product_category` | FeedBridge taxonomy normalisation outputs Google Product Taxonomy-compatible paths |
FeedBridge's taxonomy normalisation feature outputs Google Product Taxonomy-style category paths from the enrichment pipeline — making the `google_product_category` field in the GMC export usable without manual taxonomy mapping. FeedBridge's AI enrichment also produces title and description content that is optimised for discoverability, which maps directly into the GMC `title` and `description` fields in the CSV export.
---
Feed Hosting, Scheduling, and Health for GMC
Like the ACP JSON-LD feed and UCP Interactive Protocol, FeedBridge's Google Merchant Center CSV export is hosted on CDN-backed URLs unique per brand. This means the Merchant Center account can be configured to fetch the feed from a stable, hosted URL at a scheduled interval — rather than requiring manual upload of a file each time the catalog changes.
FeedBridge's feed scheduling feature supports configurable auto-refresh intervals, keeping the GMC CSV export current with catalog changes, price updates, and availability changes. Feed health monitoring runs continuous checks for dead product URLs and validity issues, alerting merchants via the feed alert preferences system when problems are detected.
For Merchant Center specifically, feed freshness matters because GMC has its own freshness requirements — products that are not updated within 30 days begin to degrade in Merchant Center's quality scoring. Scheduled feed refresh from FeedBridge's hosted URL keeps the feed current within the merchant's normal catalog update cycle, preventing freshness-related product degradation.
---
GMC CSV vs ACP JSON-LD: What Each Covers
Merchants implementing both the GMC CSV and the ACP JSON-LD feed in FeedBridge should understand what each format covers and where they complement each other:
| Dimension | GMC CSV | ACP JSON-LD | |---|---|---| | Channel | Google Shopping, Google AI Mode (via Merchant Center) | ChatGPT Shopping, ACP surfaces | | Format | CSV (tabular) | JSON-LD (structured linked data) | | AI enrichment fields | Not included in standard GMC spec | Intent tags, Q&A, personas, use cases, voice snippets | | Trust signals | Not in standard GMC spec | `accepts_returns`, `return_deadline_days`, `is_digital` | | Variant structure | `item_group_id` + variant attributes | `variant_dict` mapping | | Sale pricing | `sale_price` + `sale_price_effective_date` | `sale_price` + start/end date fields | | Live cart/checkout | Not in feed; UCP handles this for Google | Not in ACP feed; ACP checkout API handles this |
Both formats are generated from the same enriched product record in FeedBridge. The GMC CSV does not carry ACP enrichment fields (intent tags, Q&A pairs, persona arrays) — these are ACP-specific extensions beyond the GMC product data specification. However, the underlying enrichment work in FeedBridge (AI-enhanced titles, descriptions, category normalisation) improves the quality of both the GMC CSV and the ACP JSON-LD feed simultaneously.
---
Why It Matters for Merchants
Google Merchant Center remains the primary data gateway for Google Shopping and Google AI Mode. A correctly structured, complete, and current GMC CSV feed is the prerequisite for any product visibility on Google's commerce surfaces — including the AI surfaces that are increasingly how buyers discover and evaluate products. An approved, high-quality GMC feed does not guarantee visibility, but disapproved or incomplete products are definitively excluded.
For merchants already investing in AI enrichment for their ACP feed, the GMC CSV export in FeedBridge means that enrichment investment extends to Google surfaces at no additional effort. AI-enhanced titles and descriptions, normalised category paths, complete variant attributes, and current sale pricing fields all map directly into the GMC CSV export, producing a higher-quality Merchant Center feed than a basic mapping from raw product data would yield.
---
FeedBridge Relevance
FeedBridge exports Google Merchant Center CSV as a live feed format, hosted on CDN-backed URLs with configurable scheduling and feed health monitoring. The export is generated from the normalised, AI-enriched product record — mapping FeedBridge fields to GMC attribute requirements including title, description, price with currency, availability, brand, GTIN, MPN, sale pricing, and Google Product Taxonomy-style category paths from FeedBridge's taxonomy normalisation pipeline.
Auto-field mapping handles 13 common columns during catalog import, aligning source data to GMC field requirements from ingestion. Custom Feed Templates allow custom field mappings and filters to be applied to the GMC CSV output for specific Merchant Center account configurations.
The AI Readiness Score's Content Quality dimension (25%) evaluates the fields that most directly affect GMC feed quality — title length and richness, description completeness, category normalisation, and image presence — giving merchants a measurable indicator of how well their catalog data is prepared for Merchant Center submission.
---
Frequently Asked Questions
Q: Does FeedBridge submit the GMC CSV to Merchant Center directly, or do I need to register the feed URL myself? A: FeedBridge hosts the GMC CSV on a stable CDN-backed URL per brand. Registering that URL in your Merchant Center account as a scheduled fetch source is a one-time setup step performed in Merchant Center. Once registered, Merchant Center fetches the feed from the FeedBridge-hosted URL on your configured schedule — no manual file upload required after the initial registration.
Q: If I improve my product titles and descriptions in FeedBridge, does that improvement appear in the GMC CSV? A: Yes. FeedBridge's AI-enhanced title and description fields map directly to the `title` and `description` attributes in the GMC CSV export. Any enrichment improvement applied to a product in FeedBridge is reflected in the next feed refresh. This means that AI enrichment work done for ACP feed quality also improves the GMC CSV output simultaneously.
Q: Does the GMC CSV include ACP enrichment fields like intent tags and Q&A pairs? A: No. ACP enrichment fields (intent tags, persona arrays, use cases, Q&A pairs, voice snippets) are ACP-specific extensions that are not part of Google's product data specification. The GMC CSV export maps to the standard GMC attribute set. Those enrichment fields are included in the ACP JSON-LD feed and the UCP endpoint responses, which serve the AI surfaces that consume them.
Q: What happens if my GTIN is missing from the GMC CSV? A: For branded products with an assigned GTIN, missing GTIN causes Merchant Center to flag the product with a unique product identifier warning, which can result in limited visibility or disapproval depending on the product category. In FeedBridge, GTIN presence is evaluated as part of the Commerce Signals dimension of the AI Readiness Score — a missing GTIN reduces the commerce signals sub-score and is surfaced as an actionable fix suggestion in the per-product checklist.
Q: How is sale pricing formatted in the GMC CSV? A: In the GMC CSV, sale pricing uses two fields: `sale_price` (numeric with currency, e.g., `19.99 USD`) and `sale_price_effective_date` (the start and end date-times of the promotional period, formatted as an ISO 8601 range). FeedBridge maps its `sale_price`, `sale_price_start_date`, and `sale_price_end_date` fields to this GMC format in the CSV export.
---
Related Topics
Parent hub: Product Feeds and Commerce Data — Feed Formats
Related concepts:
- UCP Merchant Center Onboarding
- UCP Interactive Protocol Support
- ACP Feed JSON-LD Explained
- Hosted Feed URLs Explained
- Product Attributes That Improve Discoverability
Breadcrumb:
---
Source Documentation
| Claim | Source | Source Class | Reference | |---|---|---|---| | Google Merchant Center CSV: live feed format in FeedBridge | FeedBridge Platform Capabilities April 2026 v2.0 | T1 – FeedBridge Internal | FeedBridge-Platform-Capabilities-April2026.md | | Auto field mapping: 13 common columns including SKU, title, price, GTIN, MPN, images | FeedBridge Platform Capabilities April 2026 v2.0 | T1 – FeedBridge Internal | FeedBridge-Platform-Capabilities-April2026.md | | Hosted Feed URLs: CDN-backed, unique per brand — live | FeedBridge Platform Capabilities April 2026 v2.0 | T1 – FeedBridge Internal | FeedBridge-Platform-Capabilities-April2026.md | | Feed Scheduling: configurable auto-refresh — live | FeedBridge Platform Capabilities April 2026 v2.0 | T1 – FeedBridge Internal | FeedBridge-Platform-Capabilities-April2026.md | | Feed Health Monitoring, Feed Alert Preferences — live | FeedBridge Platform Capabilities April 2026 v2.0 | T1 – FeedBridge Internal | FeedBridge-Platform-Capabilities-April2026.md | | Custom Feed Templates: custom field mappings and filters — live | FeedBridge Platform Capabilities April 2026 v2.0 | T1 – FeedBridge Internal | FeedBridge-Platform-Capabilities-April2026.md | | Taxonomy Normalisation: Google Product Taxonomy-style paths — live | FeedBridge Platform Capabilities April 2026 v2.0 | T1 – FeedBridge Internal | FeedBridge-Platform-Capabilities-April2026.md | | GMC required fields: id, title, description, link, image_link, price, availability, condition, brand, gtin | Google Merchant Center Product Data Specification | T2 – Google | support.google.com/merchants/answer/7052112 | | item_group_id required for variant products in US, UK, DE, FR, JP, BR | Google Merchant Center Product Data Specification | T2 – Google | support.google.com/merchants/answer/7052112 | | Brand: one of three unique product identifiers alongside GTIN and MPN | Google Merchant Center Feed Attribute Reference | T2 – Google | storegrowers.com/google-merchant-center-feed-attributes |