QuantumAdsLab Logo Quantum Ads Lab
Google Shopping images: automatic AI analysis of products in Merchant Center
How Google automatically analyzes product images for Google Shopping

Google Shopping Images: Requirements and Disapprovals 2026

Summary

What you'll find in this article

  • How automatic image analysis works in Merchant Center, two distinct verification levels
  • Confirmed technical criteria: dimensions, formats, CMYK, accessible URLs
  • Editorial criteria: text, watermarks, overlays, placeholders, AI images, what the system detects
  • The new 500×500px requirement announced in 2026 with enforcement deadline January 2027
  • What is reasonably inferable but not officially documented
  • How to optimize Google Shopping images: practical techniques with real data

Google Shopping images are the most overlooked factor in catalog management, and often the most critical. Most advertisers spend weeks optimizing bids, feed structure, and negative keywords, while images are uploaded once and forgotten. The problem only surfaces when merchant center image disapproval notices start accumulating in the dashboard, and by then, products have already been invisible in shopping ads for days or weeks.

The point is that Google does more than just check that an image exists. Google Shopping operates on a visual system that automatically analyzes every image on at least two distinct levels, one technical and one editorial , using computer vision tools that have become significantly more precise in recent years. Understanding what the system checks and how it does so isn't theory: it's the difference between a catalog that runs smoothly and one that silently accumulates merchant center image disapproval flags.

How automatic image analysis works in Merchant Center

✅ Confirmed source, Information in this section comes from official Google Merchant Center documentation and the Google Product Data Specification.

When you upload an image to your product feed in Google Merchant Center, via the image link field (image_link attribute), the system starts a two-phase sequential verification process. If one phase fails, a merchant center image disapproval is triggered and subsequent phases are not executed.

Phase 1, Technical verification: The system checks that the image is reachable (publicly accessible URL, not blocked by robots.txt), that the file format is supported, that the dimensions meet the category minimums, and that the file doesn't exceed 16 MB. This phase is fully automated and generally resolves within 24–72 hours for new images or updated URLs.

Phase 2, Editorial verification: This is where computer vision comes in. The system analyzes image content to detect policy violations, overlaid text, watermarks, logos, placeholders, absent or unidentifiable product. This phase generates the most frequent and least obvious merchant center image disapproval cases.

Processing times vary significantly depending on how the image is updated. Good feed management means updating both the image URL and its content simultaneously: re-processing then happens within 24–72 hours, versus up to 6 weeks if you change the file but keep the same URL. This is one of the least documented patterns but has the greatest practical impact on how quickly disapprovals are resolved.

Google Shopping Image Requirements 2026: Confirmed Technical Criteria

Meeting google shopping requirements for images starts at the technical level. The table below summarizes every confirmed constraint, a single violation in any row is enough to trigger a merchant center image disapproval.

Criterion Minimum requirement Recommended Notes
Non-apparel dimensions 100 × 100 px 1200 × 1200 px From Jan. 2027: minimum 500×500 px
Apparel dimensions 250 × 250 px 1200 × 1200 px From Jan. 2027: minimum 500×500 px
File format JPEG, PNG, GIF, WebP, BMP, TIFF JPEG or WebP GIF must be static, not animated
File size Under 2 MB Maximum 16 MB
Color mode RGB RGB CMYK not supported, causes silent disapprovals
URL Publicly accessible Stable over time Not blocked by robots.txt, must return HTTP 200
✅ Confirmed by Google: The system can detect images that have been artificially enlarged via upscaling. An originally low-resolution image resized to minimum dimensions using resize tools will not pass quality verification. Official image_link Google Merchant Center documentation.

A common error I've seen on large e-commerce feeds is the CMYK Google Shopping error. Images exported from professional design software (Illustrator, InDesign) often use CMYK mode for print. When uploaded to Merchant Center they trigger a merchant center image disapproval that appears generic in diagnostics, often as "Poor image quality", without explicitly indicating the color profile as the cause. Converting to RGB before export eliminates this problem at the source.

Google Shopping Image Editorial Criteria: What Computer Vision Detects

This is the least documented and most interesting part. Google has confirmed that the system automatically analyzes the content of Google Shopping images, but has never published a technical description of the models used. What we know comes from policy documentation and the system's observable behavior.

Overlaid text and promotional overlays

✅ Confirmed by Google (Merchant Center documentation): The system automatically detects promotional text, watermarks, logos, and any overlay covering the product. The disapproval is flagged as "Promotional overlay on image" or "Text on image" in the Merchant Center Needs Attention section. Google offers the Automatic Image Improvements feature, which can automatically remove these elements if enabled in the account.

Concrete examples of overlays the system detects: text with prices or discount percentages burned into the image ("50% OFF", "€29.99"), promotional badges ("NEW", "SALE", "FREE SHIPPING"), brand or photographer watermarks, company logos overlaid on the product. The logic is that Google wants clean images, it adds promotional badges itself (Sale Badge, Price Drop Badge) based on feed data, not on the image.

White background in Google Shopping: mandatory or recommended?

This is one of the most frequent questions: is a white background a mandatory requirement for Google Shopping images? The answer is no, Google does not automatically disapprove images with colored backgrounds. A white background is strongly recommended because clean product images on a neutral background ensure the product is the dominant visual element and adapt to carousel templates. The actual rule is that the background must not distract from the product. Busy backgrounds, complex patterns, or elements that obscure the product are common merchant center image disapproval causes.

Product presence and identifiability

The system verifies that the product being sold is the main, visible subject of the image. Disapproval causes in this category include:

Placeholder images: any generic image such as "No Image Available," icons, logos as the main image, images showing empty packaging instead of the product. Immediate merchant center image disapproval.

Unidentifiable product: images too small for the product to be distinguishable, blurry or low-resolution images, angles that don't clearly show the product.

Product-variant mismatch: if you sell red shoes but show the black model's image, the system may flag the inconsistency. For online stores with many color variants, this is one of the hardest problems to track manually , and it can block entire Shopping campaigns without the diagnostic clearly indicating it.

Category-specific rules

✅ Confirmed by Google documentation: For apparel, products may be shown on a model in both main and additional images. For non-apparel products (shoes, bags, accessories), the main image must show the product alone , model images go in additional images via additional_image_link. For bundles, the main image must show all included items.

AI image detection in Shopping images

⚠️ Reasonably inferable, not officially documented as a mechanism: The system is becoming increasingly capable of detecting AI-generated backgrounds in Google Shopping images. Three signals are detected: IPTC metadata automatically embedded by tools like Adobe Photoshop Generative Fill or Google Product Studio, shadow and light vector inconsistencies between the real product and the artificial background, and synthetic pixel patterns that differ from real camera sensor noise. Google requires images to faithfully represent the physical product, it's not AI detection per se that causes a merchant center image disapproval, but misrepresentation.

In practice: using AI to remove the background and replace it with clean white rarely causes issues. Using AI to add artificial lifestyle backgrounds in place of real studio photography is the most problematic case, especially if the tool used embeds metadata that Google reads.

2026 Update: Google Shopping Images at 500×500 Pixels, January 2027 Deadline

⚠️ Critical 2026 update: Google updated the Product Data Specifications in April 2026. The new 500×500 pixel requirement takes effect on January 31, 2027, warnings in Merchant Center for all images below that threshold began on April 14, 2026. Source: ALM Corp analysis of the April 2026 update.

This update has potentially enormous impact on large catalogs built over the years with minimum-quality images. The previous requirement was 100×100 pixels for non-apparel, a threshold so low that many merchants paid no attention to it. The jump to 500×500 means images that were "sufficient" for years become non-compliant.

Google has stated it may apply AI upscaling to some undersized images, but explicitly specified that merchants should not rely on this as a permanent solution. The preferred path remains updating catalog images.

For catalogs with thousands of SKUs, January 2027 seems far away but it isn't. Identifying images below 500×500, tracing back to high-resolution originals, updating the feed, and waiting for re-crawl are operations that take weeks on large catalogs. Those who wait until the last moment will face massive merchant center image disapproval waves on February 1, 2027.

Do Images Influence Ranking in Google Shopping?

⚠️ Third-party data and inferences, not officially documented by Google as ranking mechanisms.

Beyond approval or disapproval, there's a question few ask: do product images influence positioning in the Shopping carousel? Google has never documented an "image quality score" that influences ranking, in Google Shopping campaigns, positioning is based on bid, feed relevance, and landing page quality, not on image aesthetics alone. However, performance data suggests an indirect correlation.

Data collected by TrustedWebEServices indicates that high-quality images (800×800+ pixels, white background) generate 27% more CTR than minimum-size images on non-neutral backgrounds. CTR is a signal Google uses to evaluate overall listing quality, a listing with higher CTR tends to achieve better positioning over time, even if Google doesn't explicitly confirm this mechanism for Shopping.

The most likely explanation is that Google Shopping image quality doesn't directly influence algorithmic ranking, but it does influence CTR, which in turn indirectly influences visibility. From a practical standpoint the result is the same: better images bring more traffic, regardless of the exact mechanism.

Google Shopping Image Optimization: Practical Guide

Target 1200×1200 pixels as your catalog standard. Not the minimum, the optimal. Google displays Shopping images on high-density screens (Retina, 4K) and on mobile. A 100×100 image that passes technical approval appears pixelated in real contexts. 1200×1200 covers all placements without compromise and also protects you from the new 500×500 enforcement.

Check the color profile before uploading. Always export in RGB, not CMYK. In Photoshop: Image → Mode → RGB Color. In Lightroom: RGB is the default, no action needed. If you're using files from a designer or agency, always verify, print deliverables are almost always CMYK and cause silent merchant center image disapproval flags.

Update the URL along with the image when fixing a problem. If an image gets a merchant center image disapproval and you replace it while keeping the same URL, re-crawling can take up to 6 weeks. Updating the URL (even just adding a parameter like ?v=2) forces re-processing in 24–72 hours. On large catalogs with widespread disapprovals, this difference equals weeks of lost visibility.

Don't use AI images with embedded metadata. If you use Photoshop Generative Fill for backgrounds, remove IPTC metadata from the final image before uploading (Photoshop: File → File Info → delete all). This doesn't solve the visual discrepancy issue if the background is obviously artificial, but it eliminates the most direct signal the system reads.

Use the additional image link attribute (additional_image_link) for lifestyle images. The main image (image_link) must follow stricter rules, clean product, neutral background, no overlays. Additional images (up to 10 via the additional image link attribute) have more permissive requirements and can show the product in context, on a model, from different angles.

Regularly audit your merchant center account Diagnostics. Merchant center image disapproval notices accumulate silently, you don't receive an email for each product. The Needs Attention section in your merchant center account is the only place where they appear aggregated. Industry data indicates that merchant center image disapproval issues account for approximately 65% of all Merchant Center rejections.

FAQ on Google Shopping Images

Why does Google Shopping disapprove images with text or watermarks?
Google uses automated computer vision systems that detect promotional text, watermarks, logos, and overlays in Google Shopping images. The disapproval appears as "Promotional overlay on image" in Merchant Center. The product is not shown in ads until resolved. Google offers the Automatic Image Improvements feature, activatable in Growth > Manage Programs, which can automatically remove these elements from the image.
Is a white background mandatory for Google Shopping images?
No, a white background is not a mandatory technical requirement for Google Shopping images, an image with a colored background is not automatically disapproved. Google strongly recommends it to ensure the product is the dominant element. The actual rule is that the background must not distract from the product. Busy backgrounds, complex patterns, or elements that obscure the product are editorial disapproval causes.
What are the minimum size requirements for Google Shopping images in 2026?
Current minimum requirements for Google Shopping images are 100×100 pixels for non-apparel products and 250×250 pixels for apparel. From April 2026, warnings began for images below 500×500 pixels, with actual enforcement (disapproval) from January 31, 2027. Google recommends 1200×1200 pixels for best performance across all devices and placements.
Does Google Shopping approve AI-generated product images?
Google's automated systems are becoming more capable of detecting AI-generated images, through IPTC metadata embedded by tools, shadow and light vector inconsistencies, and synthetic pixel patterns. Google requires images to faithfully represent the real product. Using AI to remove the background and replace it with white is generally safe. Using AI to generate artificial lifestyle backgrounds, especially with tools that embed metadata, increases the risk of disapproval for misrepresentation.
Does the CMYK color profile cause disapprovals in Google Shopping?
Yes. Google Shopping images must be in RGB color mode. CMYK files, typical print exports from Illustrator or InDesign, generate "Poor image quality" errors in Merchant Center without explicitly indicating the color profile. It's one of the most insidious problems because it's invisible to the eye. Before uploading any image, check: Photoshop → Image → Mode → should read "RGB Color".
How long does it take to reprocess an updated Google Shopping image?
24–72 hours if you update both the image URL and content. Up to 6 weeks if you keep the same URL and only change the file. The most effective strategy for resolving disapprovals quickly is to update both the file and the URL, even adding a parameter like ?v=2 forces immediate re-crawl instead of waiting for the natural Merchant Center cycle.

All articles

See all →