Google’s 500 x 500 Product Image Requirement
From 31 January 2027, every product image submitted to Google Merchant Center must be at least 500 x 500 pixels. Products whose image_link falls below that will no longer be approved and will stop appearing on Google – in Shopping ads and in free listings alike.
The requirement applies across all product categories and all marketing methods. It also applies to additional_image_link, not just your primary image.
Here is the part most coverage understates: a product sitting at 300 x 300 pixels is perfectly compliant today and disapproved the moment the cutover happens, without a single thing about it changing. Nothing in your catalogue has to go wrong for you to lose the listing.
And the enforcement date is not your working deadline. 31 January falls immediately after peak trading, in the middle of post-Christmas reconciliation, when most retail teams are still under a change freeze and nobody is commissioning photography. If your catalogue needs remediation, the realistic window closes in October 2026 – and if it needs new photography, the budget decision needs making in this financial year, not the next one.
Analysis by Konrad Szymaniak, Founder of Szymaniak Digital, an enterprise SEO and AI search consultancy. Konrad has led search and commercial search programmes for organisations including Frasers Group and Sports Direct, speaks at brightonSEO and MeasureFest, and guest lectures on enterprise AI search and agent optimisation at the University of Southampton and the University of Stirling.
Last reviewed: August 2026
What Happened? Google’s 500 x 500 Product Image Requirement
Google has raised the minimum resolution for product images submitted through the image_link and additional_image_link attributes to 500 x 500 pixels, across every product category and every marketing method. Warnings have been running in Merchant Center since spring 2026. Enforcement — meaning actual product disapproval — begins on 31 January 2027.
What the requirement replaces.
Until the cutover, two different minimums apply: 100 x 100 pixels for non-clothing products and 250 x 250 pixels for clothing. Both disappear, and apparel stops being a special case on dimensions.
That means the scale of the increase depends entirely on what you sell. Apparel merchants face a doubling in each dimension. Everyone else faces a five-fold increase in each dimension — which, because area scales quadratically, is a twenty-five-fold increase in pixel count. If you sell anything other than clothing and your catalogue has been sitting comfortably above 100 x 100 for years, this is a far bigger jump than the headline number suggests.
The other numbers that matter.
- Maximum: 64 megapixels, and no more than 16 MB in file size.
- Google’s actual recommendation: around 1500 x 1500 pixels or above, which Google states is needed for best performance across all listing formats. This is the number worth building to; 500 is the floor, not the target.
- Composition: Google recommends the product occupy no less than 75% and no more than 90% of the frame.
- YouTube Shopping ads already require 500 x 500 to serve on TV – a useful indicator of where the requirement originates.
The automatic safety net, and its catch.
Google has said it will optimise some images that fall below 500 x 500 — either by substituting a high-resolution near-duplicate image it already holds, or by AI-upscaling the original. Where this happens, the optimised image meets the new requirement and prevents disapproval without any action from the merchant. Optimised images are flagged in the Needs attention section so they can be identified.
Read the qualifiers carefully. Google says it will do this “in some cases” and “when possible”. It is not a guarantee, not a commitment, and not something a merchant can verify in advance for any given SKU.
How to find your affected products.
In Merchant Center, open the Needs attention tab, select View history, and look for two orange warnings on the right-hand side: “Image too small for upcoming enforcement” and “Image resolution optimisations applied”. The second one tells you which products Google has already stepped in on. To pull the list, click View fix, then Download impacted products.
Konrad Szymaniak, Szymaniak Digital: “The thing I’d want a commercial director to understand is that this is not a feed error you fix when it breaks. Nothing is broken. Every affected product is trading normally right now and will keep trading normally until a date in January when a switch flips and some proportion of your catalogue disappears from Google. That’s a very unusual risk profile, and it’s precisely the kind that gets deprioritised for eleven months and then becomes an emergency.”
Why Does It Matter? Google’s 500 x 500 Product Image Requirement
1. The enforcement date is the worst possible date.
31 January 2027 sits directly after peak trading. Consider what a retail team is actually doing in the weeks before it: Black Friday, Christmas, Boxing Day and January sales, followed by returns processing, reconciliation and post-peak analysis. Most e-commerce operations run a change freeze from roughly early November through to mid-January. Photography studios book up months ahead. Budget cycles close.
So the enforcement date is not the deadline. Working backwards:
- Remediation must complete by late October 2026, before change freeze.
- Photography must be commissioned by roughly August–September 2026 if new shoots are needed, allowing for scheduling, shooting, editing and feed integration.
- The budget decision needs making now, because a catalogue reshoot is a capital request that will not survive being raised in December.
A merchant reading this in August has a comfortable runway. The same merchant reading it in December has a genuine problem, because the only remaining options are the ones that compromise quality.
2. The automatic optimisation is a strategic risk dressed as a convenience.
Google stepping in to upscale your images sounds like a favour. Think about what it actually means: Google is choosing your product imagery for you — the single most influential conversion asset on a product listing — using an algorithm, without your review, on products you have not identified.
Two consequences follow. First, an AI-upscaled thumbnail satisfies a pixel-count check but does not sell a product; upscaling adds pixels, not detail, and the artefacts are visible. Second, and more awkwardly, Google’s own image quality guidelines separately penalise blurry, pixelated and poorly lit images. Passing the dimension check with an upscaled image does not exempt you from the quality check.
The commercial framing is simpler than the technical one: you are competing for attention in a visual format against merchants who shot their products properly. Compliance is not the objective. Selling is.
3. Rushing the fix with AI tools creates a second compliance problem.
This is the trap almost nobody is flagging, and it will catch teams who leave this late.
Google requires AI-generated product images to carry IPTC DigitalSourceType metadata declaring how they were made — TrainedAlgorithmicMedia for fully AI-generated images, and CompositeSynthetic for real product photographs with AI-generated backgrounds or elements. Google’s own Product Studio applies this metadata automatically. Third-party AI tools may not, and metadata is routinely stripped by image pipelines, CDNs and optimisation plugins.
So a merchant who hits December, panics, and runs their catalogue through a third-party AI upscaler has plausibly traded a dimension warning for a metadata compliance issue — and will not find out until products start being reviewed.
4. There are fewer product slots to compete for than there used to be.
Image quality has always mattered. It matters more now because the surfaces are more visual and more selective. Research published by Productrise in July 2026 found AI Mode surfacing roughly 95% fewer product listings than standard search, with standard search averaging 22.5 products per page against AI Mode’s 4.3. Fewer slots means a higher bar for the products that fill them.
The YouTube Shopping requirement points the same way: 500 x 500 has been required for TV serving for some time. Google is levelling the floor across surfaces because the surfaces themselves are getting larger and more visual — connected TV, AI-assembled shopping answers, agentic purchasing flows. A 100 x 100 image was acceptable when it rendered as a thumbnail beside a blue link. It is not acceptable when it renders on a television.
5. Undersized images hide in the least-visited parts of a catalogue.
The affected products are rarely the ones anyone is looking at. They cluster in predictable places: legacy SKUs imported during a platform migration, long-tail products nobody has touched in years, discontinued lines still listed, supplier-provided imagery from smaller vendors, and variation-level images on variable products. Auditing your top sellers will tell you almost nothing, because your top sellers were photographed properly.
The same applies to additional_image_link. The requirement covers additional images too, and secondary imagery is systematically lower priority in most merchandising workflows — which is exactly why it is more likely to be undersized.
Konrad Szymaniak, Szymaniak Digital: “Every retailer I’ve worked with has a long sku list of products that quietly earn money and that nobody has thought about since they were uploaded. It won’t show up in a review of your top hundred SKUs, and the revenue at risk is genuinely material once you add it up – which is the argument that gets the photography budget approved, not the compliance one.”
Who Is Affected? Google’s 500 x 500 Product Image Requirement
- Retailers with large or legacy catalogues – the highest exposure by volume.
The more SKUs and the longer the trading history, the more undersized images are sitting unnoticed. Merchants who have migrated platforms are especially exposed, because migrations frequently carry across whatever image assets existed rather than re-sourcing them.
- Non-apparel merchants — the largest relative jump.
Moving from a 100 x 100 minimum to 500 x 500 is a twenty-five-fold increase in required pixel count. Categories that historically tolerated small imagery — auto parts, industrial supplies, spares and components, books and media, homeware accessories — face the biggest gap between where they are and where they need to be.
- Marketplace sellers and drop-shippers.
When product imagery comes from a supplier or manufacturer rather than your own studio, remediation is a procurement conversation rather than a production one. Start it now, because supplier response times are the constraint and you cannot compress them in January.
- Apparel and fashion.
The dimensional increase is modest, from 250 x 250 to 500 x 500. But fashion competes hardest on visual quality, and Google’s 1500 x 1500 recommendation matters far more here than the 500 floor. Fashion merchants building to the minimum are solving a compliance problem while losing a commercial one.
- Small and independent merchants.
Without a studio, a feed manager or an agency, the automatic optimisation is genuinely tempting. It is also the group least able to detect when it has not been applied to a given product, and least able to react quickly if it stops.
- Enterprise and multi-market operations.
One catalogue, multiple feeds, multiple markets, possibly multiple Merchant Center accounts and different image pipelines per region. The audit needs running per account, and remediation needs sequencing so that a fix in one market does not break the feed in another.
What Should Businesses Do? Google’s 500 x 500 Product Image Requirement
1. Run the audit this month, and size the revenue at risk… not the product count.
In Merchant Center, go to Needs attention, select View history, and look for “Image too small for upcoming enforcement” and “Image resolution optimisations applied”. Click View fix, then Download impacted products to export the list. Then do the step most teams skip: join that list against your sales data. “Four thousand products affected” is a number that gets deprioritised. “Four thousand products representing £X of annual revenue” is a number that gets a budget.
2. Check additional_image_link, not just your primary images.
The requirement covers both. Secondary imagery is lower priority in most workflows and correspondingly more likely to be undersized, and it is the part of the audit most commonly missed.
3. Identify where Google has already stepped in — and treat those as unresolved.
The “Image resolution optimisations applied” warning tells you which products are currently being carried by Google’s automatic optimisation. Those products are not disapproved, so they will not appear urgent. Treat them as a list of products where Google has chosen your imagery for you, and prioritise them by commercial value rather than by warning severity.
4. Build to 1500 x 1500, not to 500 x 500.
Google’s stated recommendation is around 1500 x 1500 or above for best performance across all listing formats. If you are commissioning new photography or re-exporting from source files, there is no meaningful cost difference between exporting at the minimum and exporting well above it — and doing the latter means you will not revisit this the next time Google raises the floor. Keep within the 64-megapixel and 16 MB ceilings.
5. Re-export from source before you re-shoot.
Before committing to photography, check whether higher-resolution originals already exist. In a great many cases the small image in the feed is a downsized export, a legacy thumbnail or a CDN-transformed variant, and the original asset is sitting in a DAM or a supplier folder at full resolution. Re-exporting is faster and cheaper than shooting, and it should always be the first remediation route tested.
6. If you use AI tools, verify the metadata survives your pipeline.
Google’s Product Studio adds the required IPTC DigitalSourceType metadata automatically. Third-party tools may not, and image optimisation plugins and CDNs frequently strip metadata in transit. If you are using AI generation or upscaling anywhere in the process, confirm the declaration is present in the file Google actually fetches — not in the file your tool produced.
7. Sequence the work against your trading calendar, not against Google’s date.
Audit in August or September. Remediate through September and October. Complete before change freeze in early November. Reserve January for verification only. Build in the 24–72 hours Merchant Center needs to reflect changes on the Needs attention page, and remember that a large feed re-upload is itself a change-freeze event at most retailers.
8. While you are in there, fix the adjacent image problems.
The most common image disapproval is not size – it is promotional overlays. Any “SALE”, “FREE DELIVERY”, price badge or watermark on a primary image is a documented disapproval trigger, and the same words are perfectly fine in ad copy or a merchant promotion. If you are auditing imagery anyway, catch these in the same pass rather than running the exercise twice.
What We’re Watching Next
- Whether the automatic optimisation persists after enforcement.
Google’s current framing is that optimisation prevents disapproval during the warning period. Whether it continues as a permanent safety net after 31 January 2027, or was only ever a transition measure, has not been stated. Merchants planning to rely on it are betting on an undocumented behaviour continuing past the date it was introduced to bridge. We will update this page if Google clarifies.
- Whether 500 x 500 is a destination or a waypoint.
Google already recommends 1500 x 1500 for best performance, and the gap between a recommendation and a requirement has historically closed rather than widened. The direction of travel — connected TV shopping, AI-assembled product answers, agentic purchasing — all points toward larger and more scrutinised imagery. Building to the recommendation now is the cheaper version of doing this twice.
- How image quality signals develop in AI shopping surfaces.
Dimensional compliance is a floor. What is not yet clear is how image quality influences whether a product is selected for the small number of slots available in AI-assembled shopping responses. Google has not published anything on this, and the observational research is thin. It is a reasonable hypothesis that clean, high-resolution, well-composed imagery helps — but we would flag that as inference rather than established fact, and we will say so until there is evidence either way.
- Whether Merchant Center adds better tooling.
The current workflow — Needs attention, View history, scan for orange warnings, View fix, download — is more buried than a change of this scale warrants. Whether Google surfaces a dedicated readiness view before enforcement is worth watching, particularly for merchants without feed management tooling.
- Adjacent specification changes on the same timeline.
The image requirement arrived alongside other 2026 specification updates, including product-level shipping attributes and a video link attribute. Merchants opening their feed for image remediation have a natural opportunity to address those in the same release rather than scheduling a second round of feed work.
Konrad Szymaniak, Szymaniak Digital: “The pattern I’d point to is that Google keeps raising the floor on things that used to be optional, and every time it does, the merchants who were already building to the recommendation rather than the minimum simply don’t notice. That’s the position worth being in. Shoot at 1500, export at 1500, and this becomes a calendar entry you delete rather than a project you resource.”
Frequently Asked Questions
When exactly does the 500 x 500 requirement start?
Enforcement begins on 31 January 2027. From that date, products with images below 500 x 500 pixels in the image_link attribute will no longer be approved and will stop showing on Google. Warnings have been appearing in Merchant Center since spring 2026, and those warnings do not currently prevent products from serving – which is exactly why they are easy to ignore.
What are the current minimums, and how big is the increase?
Until enforcement, the minimum is 100 x 100 pixels for non-clothing products and 250 x 250 pixels for clothing. Both are replaced by a single 500 x 500 requirement across every category and marketing method. For apparel that is a doubling in each dimension. For everything else it is a five-fold increase per dimension – a twenty-five-fold increase in total pixel count, since area scales quadratically.
Does this apply to additional images as well?
Yes. The requirement covers both image_link and additional_image_link. Secondary imagery tends to sit lower in merchandising priorities and is therefore more likely to be undersized, so it is worth auditing deliberately rather than assuming it matches your primary images.
Google says it will optimise small images automatically. Can I rely on that?
Not safely. Google has said it will optimise some sub-500 images — using a high-resolution near-duplicate it already holds, or by AI-upscaling the original — and that this prevents disapproval without merchant action. But the language is “in some cases” and “when possible”, it is not a guarantee for any given product, and no commitment has been made that it continues after enforcement begins. There is also a commercial argument against relying on it: upscaling adds pixels rather than detail, Google’s separate image quality guidelines still apply, and it means Google is choosing your product imagery for you.
How do I find which of my products are affected?
In Merchant Center, open the Needs attention tab and select View history. On the right-hand side, look for the orange warnings “Image too small for upcoming enforcement” and “Image resolution optimisations applied”. To export the list, click View fix and then Download impacted products. Cross-reference that export against your product data to identify which image_link values need replacing – and against your sales data to understand what revenue is actually at stake.
What size should I actually be using?
Google recommends around 1500 x 1500 pixels or above for best performance across all listing formats, with maximums of 64 megapixels and 16 MB. Treat 500 x 500 as the compliance floor and 1500 x 1500 as the target. If you are re-exporting or commissioning photography, the cost difference between the two is negligible and the second option means you are unlikely to repeat this exercise.
Can I use AI to upscale or generate product images?
Yes, with a condition that catches people out. Google requires AI-generated imagery to carry IPTC DigitalSourceType metadata declaring how it was produced — TrainedAlgorithmicMedia for fully AI-generated images and CompositeSynthetic for real photographs with AI-generated backgrounds or elements. Google’s Product Studio adds this automatically; third-party tools may not, and metadata is often stripped by CDNs and image optimisation plugins. Verify the declaration exists in the file Google fetches, not just the file your tool exported. And before reaching for AI at all, check whether higher-resolution originals already exist in your DAM or supplier assets — re-exporting a real photograph beats upscaling a small one every time.
What happens to my products if I miss the deadline?
Affected products are disapproved and stop appearing on Google — both in Shopping ads and in free listings. Nothing else about the product needs to change for this to happen. Fixing it afterwards is straightforward in principle: submit compliant images and re-upload. But changes take 24–72 hours to reflect in Merchant Center, and doing catalogue-scale remediation reactively in February means losing visibility during the period you are fixing it.
About Szymaniak Digital
Szymaniak Digital is a UK-based enterprise SEO and AI search consultancy helping senior marketing and e-commerce leaders stay visible as search becomes conversational, agentic and increasingly visual.
Founded by Konrad Szymaniak, the consultancy combines technical SEO and feed quality with Generative Engine Optimisation (GEO) and Answer Engine Optimisation (AEO) – because product data quality now determines visibility across Shopping, AI Mode, AI Overviews and the emerging agentic commerce surfaces alike.
Konrad has delivered search programmes for organisations including Frasers Group and Sports Direct, holds an MSc from the University of Southampton, speaks regularly at brightonSEO and MeasureFest, and guest lectures on enterprise AI search and agent optimisation at the University of Southampton and the University of Stirling. He contributes to Semrush, Screaming Frog, Sitebulb, Majestic and Wordtracker.
Not sure how much of your catalogue is exposed? Book your specialist eCommerce SEO Consultation.
