Google is raising the minimum resolution for product images in Merchant Center to 500 x 500 pixels. The change covers the image link [image_link] and additional image link [additional_image_link] attributes across all product categories and marketing methods, and Google starts enforcing it on 31 January 2027.
A main image below the threshold gets the product disapproved. Additional images are in scope too, though the documented penalty there is softer: Google's current guidance says an undersized additional image will not be shown rather than causing a disapproval. That page predates the new minimum, so treat both as in scope.
Merchant Center has been flagging undersized images since April 2026, so if you have not opened your Needs attention tab in a while, the products at risk are already sitting there in orange.
What changes
Until 31 January 2027 | From 31 January 2027 | |
Apparel products | 250 x 250 px minimum | 500 x 500 px minimum |
All other products | 100 x 100 px minimum | 500 x 500 px minimum |
Google's recommendation | 1500 x 1500 px or above | 1500 x 1500 px or above |
Upper limits | 64 megapixels, 16 MB | 64 megapixels, 16 MB |
Apparel loses its separate threshold and every category lands on the same number. The 500 x 500 requirement is a floor, not a target. Google still recommends 1500 x 1500 or above for the best rendering across Shopping surfaces, and YouTube Shopping ads already need 500 x 500 to serve on TV.
The timeline Google published
- 14 April 2026: Google announced the requirement in the annual Merchant Center product data specification update and began issuing warnings for images that miss the new minimum.
- July 2026: Merchant Center extended those warnings to images that clear the old minimum but fall below 500 x 500.
- 31 January 2027: Enforcement begins and disapprovals follow.
Nothing in your feed has to change for a product to break. A 300 x 300 image that Google approves in December gets disapproved in February with the same URL, the same file and the same product data.
How to find the products at risk
- Open Products in Merchant Center and select the Needs attention tab.
- Select View history and review the orange warnings on the right of the page.
- Look for two labels: Image too small for upcoming enforcement and Image resolution optimizations applied.
- Select View fix, then Download impacted products to export the list as a CSV.
That CSV becomes your working list. Sort it by supplier, brand or category before you decide on a fix, since undersized images tend to cluster around one data source rather than scatter across the catalogue.
Or run the report inside Feedoptimise
Connect your Merchant Center account to Feedoptimise and the platform imports your diagnostic data alongside your product data. From there you can filter your catalogue by the image warnings, build a report around them, and see the affected products next to the attributes you would use to segment them, such as brand, category, supplier or margin.
The advantage over the CSV export is that the list stays live. Once you apply a fix, the same report tells you how many products remain undersized on the next sync, so you can track progress toward the deadline without downloading a new file every week. The same connection lets you scope a rule to those products alone, which keeps your image processing costs tied to the products that need the work.
Google may rescue some of these images, but do not plan around it
Google says it may serve an optimised version of images below 500 x 500 where possible, which can prevent a disapproval without any action from you. The Image resolution optimizations applied warning shows which products currently benefit.
Google has not published the criteria behind that optimisation, and the behaviour can change before enforcement. Treat it as a cushion for a small number of SKUs, then fix the images you can control.
Two ways to fix undersized images in bulk at feed level
Replacing images at source gives you the cleanest result. Reshoots cost money, supplier images arrive at whatever size the supplier decided, and a PIM change usually takes a queue position behind everything else your development team has committed to.
If you cannot change the images in your store, you can fix them in the feed instead, between your source data and Merchant Center. Feedoptimise offers two routes.
1. Upscale the image with a generative model
Feedoptimise offers prompt-based image modifiers that send your existing product image to a generative model, then save the result to your own Google Cloud Storage bucket. You can run them on Google Gemini, the model widely known as Nano Banana, or on OpenAI's ChatGPT Images 2.0, so you can compare output from both against your own catalogue and pick whichever handles your product type better. A prompt as short as Upscale this image. handles the resolution problem, and your original stays in place, both as a fallback if a run fails and as a way back if an edit gets published and then looks wrong.
Generative upscaling rebuilds detail rather than stretching pixels, so a 200 x 200 supplier photo can come back at a usable size without the soft edges that bicubic resizing produces. Cost is visible before you commit to the run. Each call reports its token usage and an estimated price, so multiply that figure by the number of SKUs on your impacted list and you have the budget for the whole job. In practice it costs a few cents an image.
Two things to check before you run a full catalogue:
- Quality control on a sample first. Generative models can shift logos, stitching, textures or proportions. Google disapproves images that misrepresent the product, so review 20 or 30 outputs against the originals before you scale the job.
- Keep the AI metadata intact. Google requires images created with generative AI to carry IPTC DigitalSourceType metadata, such as TrainedAlgorithmicMedia or CompositeSynthetic. A generative upscale sits inside that definition, so preserve whatever metadata the model writes instead of stripping it during processing. If you are not sure what your pipeline does, open a finished file in an EXIF/IPTC reader and confirm.
Full walkthrough with prompts and folder structure: 10 Ideas to Enrich and Scale AI Product Image Edits in Your Feeds using Feedoptimise & Google Gemini
2. Place the image on a 500 x 500 canvas
The Feedoptimise Image Editor takes a different approach. Instead of altering the product photo, you drop it onto a canvas of at least 500 x 500 pixels using a template. Safe placeholders position the image without cropping or overlap, and auto-scaling preserves the key elements of the shot. The pixel dimensions clear the requirement and the product itself stays untouched, which removes the quality control burden that comes with generative edits.
For Google Shopping, keep this template plain: a blank canvas, your product image centred on it, nothing else. The Image Editor supports badges, price overlays and dynamic promotional messaging, and those belong on channels that allow them. Google disapproves feed images carrying promotional elements, so the template you build for this job should contain no text, no logos and no borders.
The trade-off is framing. Google recommends that the product fill between 75% and 90% of the image, and padding works against that:
- A 400 x 400 image on a 500 x 500 canvas fills 80% of the frame, comfortably inside the recommendation.
- A 375 x 375 image fills 75%, the lower edge of it.
- A 250 x 250 apparel image fills 50%, which passes the pixel requirement but looks like a small product floating in white space next to competitors.
Backgrounds matter too. A product shot on white pads onto a white canvas without a visible seam. A lifestyle shot with a coloured or textured background will show a clear frame around it, and Google's guidance tells you to avoid borders. The Image Editor handles this with background removal, so you can strip the original background before placing the product on a solid white canvas. That removes the seam and moves the image closer to Google's recommendation of a plain background at the same time. Quality control still applies. Every image in this workflow is undersized by definition, and fine detail like hair, fur or transparent packaging gives edge detection less to work with at low resolution, so check a sample before you run the catalogue.
Pricing runs on monthly resource packages, so you pay for the images you generate and the bandwidth you use. If you exceed your capacity, the system serves your regular image instead, which protects the feed from breaking but also leaves those products undersized. Size your package to cover the SKUs on your affected list.
Which route fits your catalogue
Use the shorter side of the source image as the deciding factor.
- 375 px or more: the canvas route gets you compliant in one pass with no risk to the product itself.
- Below 375 px: upscale with Gemini or ChatGPT Images 2.0 first. You can then run the result through a canvas template if you want a consistent square ratio across the feed.
Both modifiers can be scoped with filters, so if you have connected Google Merchant Center to Feedoptimise you can target the products carrying either image warning and leave the rest of the catalogue untouched.
Before you push the new images
- Aim for 1500 x 1500, not 500 x 500. Building to the floor means revisiting this the next time Google moves it.
- Submit a new, unique image URL. Google recrawls a new URL within about three days. If you change the file behind an existing URL, detection can take up to six weeks. Feed-level image hosting produces new URLs by default, which works in your favour here.
- Cover additional_image_link as well. The minimum applies there too, and the softer penalty on additional images is not something to build a plan around.
- Keep promotional overlays out of Google feeds. Sale badges, price tags and dynamic promotional messaging breach Google's image policy for both image_link and additional images. If you already run promotional templates for Meta or other channels, build a separate clean version for your Google Shopping feed.
- Re-check Needs attention after 24 to 72 hours. Merchant Center takes that long to reflect changes.
Getting it done before the deadline
Five months sounds comfortable until you find that half your supplier images sit below the threshold. Pull the list this week, either from Merchant Center directly or through your Feedoptimise connection, count what is affected, then decide whether upscaling, canvas placement or a mix of the two suits your catalogue.
If you want help scoping the work, book a live demo or get in touch and we will look at your feed with you.
