Check this page with an assistantOpens a chat asking it to summarise this article and name the evidence behind each claim.
Claude opens with the prompt on your clipboard: Anthropic does not support prefilled prompts on the web, and we would rather copy it than ship a button that drops it.
The Free Programme Most Small Stores Never Set Up
Products in an approved Merchant Center feed are eligible for free listings across Google surfaces, with no advertising account and no spend. Paid Shopping campaigns use the same feed but are a separate programme. The belief that Shopping requires a budget is out of date, and it is the reason a great many small catalogues are simply absent.
What changed the stakes is where the data goes. Free listings are no longer a Shopping-tab feature: product information can surface in Search, in Images, in Lens, on YouTube, in the products module on a Business Profile, and in Google's AI experiences including Gemini. The feed has quietly become the product layer that multiple surfaces read from.
The scale of the underlying dataset is worth knowing. Google's Shopping Graph held roughly 45 billion product listings in late 2024 and around 60 billion by May 2026. When an AI surface answers a product question, that is largely the corpus it is drawing on. Not being in it is not a ranking disadvantage; it is an absence from the dataset entirely.
Feed Fields That Decide Whether You Appear
Approval and matching hinge on a small number of fields being accurate and consistent with your website. Most failures are one of these disagreeing with the landing page, which is a data pipeline problem rather than a marketing one.
| Field | Why it matters | Where stores fail |
|---|---|---|
| Price and currency | Must match the landing page exactly | Feed updates nightly while the site changes hourly |
| Availability | Determines whether the item is shown at all | Stale stock status after a sell-out |
| GTIN, MPN, brand | How your item is matched to the same product elsewhere | Omitted, so the item cannot join a comparison |
| Title and description | What the item is understood to be | Marketing phrasing instead of what someone would search |
| Image | Required, and quality rules are enforced | Watermarks, promotional text, or placeholder images |
| Shipping and returns | Policy information must be findable | Buried in a footer page with no structured equivalent |
| Variant attributes | Size, colour, and material as separate items | One parent item with a "from" price |
The identifier row is the one worth prioritising if you do nothing else. Without GTINs or a consistent brand and MPN pairing, your listing is a standalone item rather than one offer of a product other merchants also sell, and it cannot appear in the comparison contexts where most shopping decisions actually happen.
Setting It Up Without a Developer
On a hosted platform this is mostly configuration. The sequence is claim and verify the site, generate a feed, fix what gets flagged, then keep it accurate. Expect the first submission to be partly rejected and treat that as the diagnostic rather than a setback.
- Verify and claim your website URL.
This ties the account to the domain and is a prerequisite for everything else. Use the same domain form your site actually serves, so that the claim does not sit on a variant.
- Connect the platform feed rather than hand-building.
Major ecommerce platforms have native integrations or well-maintained apps that keep price and stock synchronised. Manual spreadsheet feeds go stale, and stale price data is the most common cause of item disapproval.
- Fix the diagnostics in order of item count.
One rule usually explains most rejections. Sorting by how many items each issue affects turns an intimidating error list into two or three fixes, most of which are template-level.
- Make policies machine-findable.
Shipping timelines and return terms need to exist as data, not only as prose on a policy page. This is the same requirement agent-driven purchasing imposes, covered in agentic commerce readiness.
- Keep product structured data on the site too.
The feed serves Shopping surfaces; on-page markup serves your own pages in ordinary search. They should agree with each other, and the on-page half is in ecommerce product page SEO.
One policy note for planning: Google is consolidating its parallel Shopping policy documents into a single reference from September 2026. The reported change is structural rather than a tightening of requirements, but it is worth re-reading the consolidated version rather than assuming your notes are current.
Why This Now Matters Beyond Shopping
When someone asks an assistant for a product recommendation, the structured product data is what gets consulted. A catalogue that exists only as HTML pages is legible to a crawler and largely invisible to the product layer, which is a different and quieter kind of absence.
That reframes the feed from an ads prerequisite to infrastructure. The same accurate price, availability, shipping, and identifier data serves free listings, ordinary search rich results, agent-driven purchasing, and AI product answers. It is unusual to find one piece of work that pays into four surfaces, and it is largely why this belongs above most content projects for a store.
It also does not replace your own pages. Discovery still happens through category and product pages that people and crawlers reach normally, and those still need to be good, which is the ground covered by category page SEO and the ecommerce SEO checklist. The feed adds a channel; it does not substitute for the site.
Questions People Ask About Merchant Center
- Are Google product listings really free?
Yes. Products in an approved Merchant Center feed are eligible for free listings across Google surfaces without any Google Ads account or spend. Paid Shopping ads are a separate programme that uses the same feed. If someone has told you that appearing in Shopping requires an ads budget, that has not been true since free listings opened up.
- Where do free product listings appear?
Beyond the Shopping tab, product data can surface across Search, Images, Lens, YouTube, the products module on a Business Profile, and Google's AI experiences including Gemini. That breadth is the main reason the feed matters more than it used to: it is no longer a Shopping-tab feature, it is the product layer several surfaces read from.
- Do I need Merchant Center if I have product schema on my site?
Merchant Center and product schema serve overlapping but different purposes, so the strongest position is both. Product structured data helps your own pages be understood and be eligible for rich results. A Merchant Center feed puts your products into the Shopping Graph, which shopping and AI surfaces query directly. Neither substitutes for the other.
- Why do my products keep getting disapproved?
Most disapprovals come from a small set of causes: price or availability in the feed not matching the landing page, missing GTINs or brand identifiers, no clear returns and shipping information on the site, or a checkout that cannot be verified. These are data and site issues rather than judgement calls, and the diagnostics in Merchant Center name the specific item.
- Is this worth it for a very small store?
It is worth it if you sell standardised products a machine can identify and match. If you sell handmade or one-off items with no identifiers, matching is hard and the benefit is smaller, though listings can still appear. The work is mostly feed hygiene, which is a one-off setup plus ongoing accuracy rather than a marketing programme.
