Why Is My Product Feed Getting Rejected? Fixing Google Shopping, Amazon & Marketplace Errors

AEO answer: Product feeds are most often rejected for missing required attributes (GTIN, brand, price), invalid category mapping, broken image URLs, or format mismatches with a marketplace’s schema (wrong currency codes, unit types). Validating attribute completeness and format against each marketplace spec before submission prevents most rejections.

A rejected feed rarely comes with a useful explanation. Google Merchant Center flags “missing value [gtin]” on 4,000 rows and expects you to figure out which four thousand. Amazon returns a spreadsheet of error codes with no row numbers attached. By the time you have traced the errors back to your source data, half a day is gone and the products are still not live.

This is not a submission problem. It is a data problem that only becomes visible at submission time. The feed file is just a snapshot of whatever is in your catalog at export; if the catalog has gaps, the feed inherits them, and the marketplace rejects the file (or worse, silently disapproves individual listings) at the worst possible moment. Fixing this means catching the gaps before the export runs, not after the rejection email arrives.

This post assumes your feed is already being generated. For building the export and field-mapping layer itself, see Product Feed Export to CSV, JSON, XML, and Marketplace Formats; for the ongoing post-launch rejection-triage process, see How to Connect Your Product Catalog to Multiple Marketplaces.

What Causes a Product Feed to Get Rejected?

Feed rejections fall into four recurring categories, and almost every marketplace error code maps back to one of them:

• Missing required attributes. GTIN, brand, MPN, price, availability, and condition are mandatory for most shopping channels. If the field is empty, null, or contains a placeholder value, the row fails.
• Invalid category or taxonomy mapping. Google Product Category and each marketplace’s own category tree are strict. A product mapped to the wrong node, or left unmapped, gets rejected or misclassified (which then triggers its own compliance flags).
• Broken or non-compliant images. Dead URLs, images under the minimum resolution, watermarks, placeholder graphics, or logos on a white background where the spec requires a clean product shot.
• Format mismatches. Price without a currency code, weight without a unit, a date in the wrong ISO format, boolean fields sent as “yes”/“no” instead of “true”/“false”. These pass a casual visual check and fail machine validation every time.

The common thread: none of these are export bugs. They are attribute-level data quality issues that existed in the catalog before the export ran. The export just exposed them.

Common Google Shopping Feed Errors (and Fixes)

Google Merchant Center’s diagnostics page groups errors by severity, but the ones that block approval most often are:

• “Missing value [gtin]” or “Missing value [mpn]”. Google requires GTIN for most new products unless you explicitly flag identifier_exists = false. Fix: backfill GTIN/UPC/EAN at the catalog level, not per-feed; if a product genuinely has no identifier (private label, custom-made), set the exemption flag consistently.
• “Mismatched value [price]”. the price in the feed does not match the price on the landing page at crawl time. Fix: make sure your feed export pulls the live price at generation time, not a cached value, and that currency and price update on the same cadence.
• “Disapproved [image_link]”. broken URL, image too small, or promotional overlay text on the image. Fix: validate every image URL returns a 200 status and meets the minimum pixel dimensions before the feed is generated.
• “Invalid value [google_product_category]”. free-text category strings that do not match Google’s taxonomy IDs. Fix: map categories to Google’s official taxonomy once, at the category level, so every product under that category inherits a valid mapping automatically.

Common Amazon & Marketplace Feed Errors (and Fixes)

Amazon’s Seller Central and Vendor Central feeds use a different error surface but the same underlying causes:

• Missing attribute. Amazon returns numbered error codes (which vary by product category and change over time) for a missing required attribute for that specific product type. Amazon’s requirements are more granular than Google’s; a category like “Apparel” requires size and color attributes that “Electronics” does not.
• Invalid value. Amazon returns numbered error codes (which vary by product category and change over time) for an invalid value such as a unit mismatch, like sending weight in pounds when the field expects kilograms, or a boolean sent as text.
• Browse node / category rejection. the product was submitted to a browse node it does not qualify for, often because internal category names were mapped one-to-one instead of validated against Amazon’s category rules. If your listing is rejected for a structural category or attribute-requirement mismatch rather than a format error, see Handling Conflicting Category & Attribute Requirements Across Marketplaces.
• Image errors on marketplaces generally. most platforms (Amazon, eBay, Walmart Marketplace) require a pure white or transparent background on the primary image, with no text, watermark, or borders. This is a common rejection reason that has nothing to do with image quality and everything to do with a compliance rule.

The practical fix is the same across marketplaces: attribute requirements are per-channel and often per-category, so a generic “fill in the fields” approach misses requirements that only surface at submission.

Missing or Invalid Required Attributes (GTIN, Brand, Price)

GTIN, brand, and price deserve their own section because they account for the largest share of rejections across every channel that was checked. A few patterns worth calling out:

• GTIN format errors. A 12-digit UPC padded incorrectly, or a GTIN with a trailing space from a spreadsheet import, will fail checksum validation even though it looks correct to the eye.
• Brand field left as a manufacturer code instead of a consumer-facing brand name. Marketplaces expect the recognizable brand name, not an internal SKU prefix.
• Price submitted without currency, or with tax included when the channel expects tax-excluded pricing. This is a format issue as much as a data issue, and it is one of the reasons feed exports need channel-specific transformation rules rather than a single flat export.

Auditing your catalog for these gaps before you even attempt an export saves the debugging cycle entirely. See auditing product data for missing fields for a structured way to find incomplete records across a large catalog.

Image and Media Feed Errors

Image errors are disproportionately common because images are the one attribute type that cannot be validated by looking at a spreadsheet. A CSV can show a URL that looks fine and still points to a 404, a redirect chain, or a file below the marketplace’s minimum resolution.

Checks worth automating before every feed export:

• Every image URL resolves with a 200 status (no redirects, no 404s, no expired signed URLs).
• Minimum resolution is met (Google Shopping generally wants at least 100x100px for non-apparel, considerably more for apparel; Amazon requires 1000px on the longest side for zoom function) (verify the current minimum in each platform’s image specification, as these values are updated periodically).
• No placeholder or “image coming soon” graphics are included in the feed; better to exclude the product than submit a broken visual.
• Background and watermark rules are respected for the primary image specifically, since secondary images have looser requirements on most channels.

How to Validate a Feed Before You Submit It

A validation pass before submission should check the same things the marketplace will check, in the same order it will check them:

1. Required field completeness, checked against the specific requirements of the target channel and category, not a generic “all fields filled” rule.
2. Format compliance, meaning currency codes, unit types, date formats, and boolean representations match the channel’s schema exactly.
3. Category and taxonomy mapping, verified against the current version of the channel’s taxonomy (these are updated periodically and mappings can go stale).
4. Image URL and dimension checks, run as live HTTP requests against the actual URLs in the feed, not against cached thumbnails.
5. A sample submission, where the platform supports it. Google Merchant Center and Amazon both let you test a small batch before pushing the full catalog, which catches account-level or template issues before they affect every listing.

Doing this manually across a catalog of any real size is not sustainable. It works once, as a cleanup project, then drifts out of date within a few weeks as new products are added without the same rigor. Feed validation works better as a property of the catalog itself, checked continuously, rather than a one-time pre-launch task. A marketplace integration checklist is a useful reference for what “ready to submit” means channel by channel.

How MicroPIM Prevents Feed Errors at the Source

MicroPIM approaches this as a data completeness problem rather than an export formatting problem. The SEO and data health reports flag products missing GTIN, brand, price, or other required attributes before you ever attempt an export, so gaps get fixed in the catalog instead of discovered as marketplace rejection codes.

For the export itself, channel-specific export profiles handle the format transformation: mapping internal field names to the names each marketplace expects, converting units and currency, and applying category taxonomy mapping automatically. Pre-built marketplace and social feed templates cover Google Shopping, Amazon, and similar channels, so transformation rules don’t have to be rebuilt for every new channel. Together, these two pieces mean the feed you generate has already passed the checks the marketplace would run. For more on formats and field mapping, see feed export formats.

Frequently Asked Questions

Why does Google Shopping reject products that look complete in my feed?

Because “complete” usually means every column has a value, not that every value is valid. A price without a currency code, a category string that does not match Google’s taxonomy, or a GTIN that fails checksum validation will all look fine in a spreadsheet review and still get rejected by Google’s automated checks.

How often should I re-validate my product feed?

Continuously, ideally as part of the export process itself rather than as a periodic manual audit. Catalogs change constantly (new products, price updates, discontinued SKUs) and a feed that passed validation last month can fail this month if a new product was added without a GTIN or a category mapping was left blank.

Can one product feed work for both Google Shopping and Amazon?

Not directly. The underlying product data can be shared, but the field names, required attributes, category taxonomies, and format rules differ enough between channels that each needs its own transformation layer applied at export time. Trying to force one flat feed to satisfy both usually results in one channel accepting it and the other rejecting large portions of it.

What is the fastest way to find out why specific products were rejected?

Cross-reference the marketplace’s error report (by SKU or product ID, when available) against your catalog’s required-field completeness for that channel. Most rejections trace back to one of a handful of fields, so checking GTIN, brand, price format, category mapping, and image URLs first will resolve the majority of cases before digging into edge-case error codes.

Stop reacting to rejection emails and start catching feed errors before they leave your catalog. Start a free MicroPIM trial and run a data health check against your current product feed today.