Product feed vs structured data: how Merchant Center and your pages work together

What is the difference between a product feed and structured data?

A product feed is a file or data connection that you submit to Merchant Center, listing your products with attributes such as title, price, availability and identifiers. Structured data is markup on your web pages, usually JSON-LD, that describes the product shown on that page. One is delivered to Google, the other is discovered by crawling.

How does the feed work?

You provide product data through a file, a scheduled fetch, an API or a platform integration. Merchant Center processes it, checks it against product data requirements and reports issues. Because you control the submission, you can update prices and availability on a schedule and add attributes specific to shopping surfaces.

What do feeds control well?

Product-level attributes in bulk, update timing, shipping and tax settings in your account, and diagnostics at item level that show exactly what Google received.

How does structured data work?

Google reads the markup when it crawls a page, using it to understand the product and, where eligible, to show merchant listing or product snippet features. The data lives on the page that visitors see, which makes it easier to compare with the visible content. Google's guide to ecommerce structured data, last updated December 10, 2025, links to Merchant Center documentation about setting up structured data for improved participation in shopping experiences.

Does markup replace a feed?

Not necessarily. Google's documentation treats them as different routes, and the best choice depends on your catalog size, platform and goals. Check the current Merchant Center help for the options available to your account.

Why do both matter for visibility?

The feed supports shopping experiences, and the markup helps Search understand and present your pages. Our article on AI shopping, feeds and structured data explains why consistent data matters for newer shopping and AI experiences.

Which one should you start with?

Get your product pages and markup correct first, because they are public and support Search. Add or refine a feed when you want deeper control over shopping surfaces.

How do feeds and markup compare side by side?

The comparison below shows where each approach is strongest. It is a planning aid, and the current Google documentation decides eligibility.

Aspect	Merchant Center feed	On-page structured data
How Google gets the data	You submit it	Google crawls pages
Update speed	Scheduled, usually under your control	Depends on recrawl
Best for	Large catalogs, bulk attributes, shopping surfaces	Search understanding and page-level features
Diagnostics	Item-level issues in Merchant Center	Rich Results Test and Search Console reports
Typical failure	Disapproval for missing or wrong attributes	Invalid or mismatched markup
Source of truth risk	Feed out of date versus site	Markup out of date versus visible page

What do the failures have in common?

Stale data. Whether it is the feed or the markup, the problem is usually that the price or stock changed on the page and not in the output. A single source of truth prevents it.

How do you create a single source of truth?

Generate the page, the markup and the feed from the same product database or platform fields, so that a change appears everywhere.

Which attributes must match across feed, markup and page?

Price, currency, availability, product identifiers, title, images, condition, brand and variant details should agree everywhere. Google's ecommerce guidance and Merchant Center policies emphasize accuracy, and mismatches lead to disapprovals or lost features.

What are the most common mismatches?

Prices that include or exclude tax differently, sale prices shown on the page but not in the feed, items marked in stock in the feed but sold out on the page, identifiers such as GTIN or MPN missing from one source, variant URLs that show a different size or color than the feed describes and images that differ from the product shown.

Why does tax treatment cause so many problems?

Regions treat displayed prices differently. Make sure the price in the feed follows the rules for your target country and matches what the shopper sees at that stage of the journey.

How do variants add complexity?

Each variant should have its own identifiers, price and availability where they differ, and its page should make the selected variant clear. Our JavaScript filters article explains how variant selection can hide content from crawlers.

Should variants have separate URLs?

It depends on your platform and users. Whatever you choose, make the relationship clear with consistent canonicals and ensure each variant that you submit can be identified on the page.

What about identifiers?

Use real GTINs, MPNs and brand names where they exist. Do not invent identifiers. When products have no identifiers, follow the guidance for that case, and do not copy from other products.

Where do identifiers belong in markup?

In the Product properties such as gtin, mpn and brand, as described in Google's product structured data documentation.

How do shipping and return policies fit in?

Return and shipping information can be provided in Merchant Center settings and in markup. Google's Organization documentation describes properties for return and shipping policies at the business level, and says merchant listing markup can override them for specific products.

Why does policy data matter?

Shoppers weigh shipping cost, delivery time and returns heavily when choosing. Clear, consistent policy data helps Google display accurate information and builds trust. Differences between the policy on your site, in Merchant Center and in markup can confuse shoppers and trigger issues.

What should the site itself show?

A dedicated returns page and a shipping page that match what you submit, linked from product pages and the footer.

How do you use Organization markup for this?

Add Organization markup to your home page with your name, logo, contact details and, where applicable, return and shipping policy properties. Keep it consistent with your visible policies. The Organization page, updated September 8, 2026, says no properties are required, so add what is accurate and useful. Our GEO versus SEO guide explains why consistent entity information helps.

Can product-level policies differ?

Yes, where products have different terms, and the product markup can override the business-level default.

What mistakes occur with policies?

Policies that differ between pages, return windows described one way in the feed and another on the site, free shipping claims without conditions and hidden fees at checkout. Align all sources before submitting.

How do you diagnose disapprovals and missing features?

Start with the diagnostics in Merchant Center and Search Console, then compare the flagged item with its page and markup. Most issues trace back to a mismatch, a missing attribute or a policy violation.

Which reports help?

Merchant Center's diagnostics show item-level issues, warnings and account-level problems. Search Console's enhancement reports show markup errors for products and merchant listings. The Rich Results Test validates a specific URL. Use them together to see both sides.

What is the difference between an error and a warning?

Errors typically prevent an item from appearing, while warnings flag missing recommended data. Fix errors first and warnings where they improve your presentation.

What is a good diagnostic routine?

1. Export the list of issues and group them by type.
2. Pick one representative item per group.
3. Compare its feed values, page content and markup side by side.
4. Identify the cause, such as stale price or missing identifier.
5. Fix at the source and wait for reprocessing.
6. Re-check and confirm that the whole group clears.

How long do fixes take to show?

Feed processing may take hours to days, and recrawling pages for markup can take longer. Record the date of each fix.

When is the problem the site, not the feed?

If Google cannot crawl or see product details, for example because of blocked resources or JavaScript-only content, markup and page checks may fail. Our Googlebot and DOM guide shows how to test this.

What should Product markup include on a product page?

Product markup should describe the one product the page is about, with a name, images, description, brand, identifiers and offer details that match the visible page. Google's Product structured data documentation lists the properties and which ones are required or recommended.

Which properties matter most?

The name, image, description, brand and identifiers such as gtin or mpn describe the product. The offers property carries price, priceCurrency, availability and, where relevant, the item condition and price valid until date. Review and aggregate rating properties can be included only when the reviews are genuine, visible on the page and about that product. Google's review snippet guidance sets the rules.

Can you mark up reviews you copied from elsewhere?

You should not mark up content that is not actually on the page or that does not meet Google's review guidelines. Self-serving or invented ratings risk losing the feature.

How should variants appear in markup?

Each variant offer should describe the actual price and availability of that variant, and the page should make clear which variant is selected. Markup that lists a lowest price for a range while the page shows only one choice can look inconsistent. Align markup with what the visitor can see and buy.

What about pages that list many products?

Category and listing pages do not describe a single product, so Product markup belongs on the individual product pages. Listing pages can use other approaches, such as clear links and descriptive text.

How do you keep markup valid?

Generate JSON-LD from the same data as the page, validate with the Rich Results Test, and monitor the merchant listing and product snippet reports in Search Console. Validation proves the syntax, and a human check proves that the content matches the page.

Which mistakes appear most?

Missing price or currency, availability values that do not match stock, duplicate Product blocks on one page, markup injected by scripts that fail and reviews that are not visible.

How do you build a high-quality product feed?

Build the feed from clean source data, follow Merchant Center's attribute requirements, keep titles and descriptions accurate and refresh price and availability regularly. Quality at the source is cheaper than fixing diagnostics later.

What does good feed data look like?

Titles that describe the product plainly, with brand, type and key attributes such as size or color, and without promotional text. Descriptions that reflect the page. Image links that load and show the product clearly. Correct categories, identifiers, condition, availability and price. Consistent landing page links that lead to the exact product.

Should titles be optimized for search?

Make them clear and specific so shoppers and systems understand the product. Avoid stuffing keywords or adding claims that the product page does not support.

How often should the feed update?

Often enough that price and availability stay accurate. For stores whose stock or prices change during the day, frequent updates or API-based changes help prevent mismatches. For stable catalogs, daily may be enough.

What happens when an item sells out?

Update availability promptly in the feed and the markup. Showing sold-out items as available leads to poor experiences and potential disapprovals.

How do you handle custom attributes and segments?

Use custom labels or similar fields to group products by margin, season or performance, then use those groups in reporting. Keep a data dictionary so everyone knows what the fields mean. Our guide to category page content shows how the same product groupings can be reflected on the site.

Who maintains the dictionary?

The data owner, with changes reviewed whenever new fields are added.

Which mistakes cost ecommerce sites the most?

The costliest mistakes are the quiet ones: stale data, inconsistent policies and duplicated sources. They rarely produce an obvious error, but they reduce eligibility and trust over time.

What are the top mistakes?

• Maintaining the feed, markup and page text by hand in separate places.
• Letting sale prices appear only on the page.
• Using the same identifiers for different variants.
• Describing return policies differently in the feed, on the site and in markup.
• Blocking product pages or images from crawlers.
• Ignoring warnings until they become errors.

Which of these is easiest to fix?

Sale prices, because they usually require moving a field into the database and reading it in every output.

How do you prevent regressions?

Add automated checks that fetch a sample of product pages, parse the markup and compare it with the feed. Alert on differences. Run the check after each deployment and each bulk price change.

What if you cannot automate yet?

Start with a weekly manual sample of ten products across categories, recording price, availability, identifiers and policy text in a sheet. Even a small routine finds recurring causes.

How do you handle supplier data?

Treat supplier feeds as inputs to be cleaned, not copied. Check identifiers, remove duplicates, rewrite descriptions so that your pages are not copies of everyone else's and verify images. This also supports your page quality, as our article on scaled content abuse explains.

Why not just copy manufacturer descriptions?

Copied text adds little value and appears on many sites. Original details, photographs and helpful guidance give shoppers and search systems a reason to prefer your page.

How do you measure the impact of feed and markup improvements?

Track approval rates, the number of items with errors, impressions and clicks for product pages and conversion rates, and compare periods with a log of changes. Improvements in data quality often show up first in diagnostics and later in traffic.

Which metrics belong on the dashboard?

Share of approved items, count of errors and warnings, Search Console impressions and clicks for product URLs, merchant listing report status, revenue from organic and free listings where available and the number of mismatches found in sampling. Use a monthly review, and record each change.

Can you attribute revenue to a single fix?

Rarely. Many factors affect sales, so use trends and consistent comparisons and avoid overstating cause and effect.

What does success look like?

Fewer disapprovals, stable markup reports, accurate prices in results and gradual growth in qualified traffic. A well-run catalog also reduces support issues from customers who found a different price or stock status in search than on the site.

When do you revisit the setup?

After platform changes, catalog expansions, new markets and when Google updates documentation relevant to your products.

How does this connect with AI shopping?

Newer shopping experiences draw on product data from feeds and pages. Accurate, complete and consistent data gives those systems what they need, which is the same work described here.

How do you set up a process that keeps data accurate?

Assign ownership, automate generation from one source, test on every release and monitor diagnostics weekly. Process matters because product data changes daily and manual edits drift.

What roles are needed?

A data owner responsible for product fields, a developer responsible for templates and feed generation, and an SEO or marketing owner who watches diagnostics and policy changes. In small shops one person may hold all three roles, and the checklist still applies.

What belongs on the weekly check?

New disapprovals, drops in approved items, markup errors in Search Console, price and stock samples compared against the site and changes in Google's documentation.

How do you test before releases?

Include markup validation, feed generation and a sample of product pages in your release checklist. Compare outputs for a set of reference products covering variants, sale prices and out-of-stock items. Catch problems before shoppers and Google do.

What if a platform plugin generates the data?

Check its output regularly. Plugins update, and defaults may not match your catalog. Google notes that an integrated extension or plugin can make adding structured data easier, but you remain responsible for accuracy.

How do you stay aware of changes?

Follow Merchant Center announcements and Google Search Central documentation, and note each page's update date. Specifications change over time, so confirm requirements against current pages and not memory.

How do platforms and plugins affect the setup?

Most stores run on an ecommerce platform with built-in or plugin-based support for feeds and markup. These tools save time, but defaults vary, so you need to verify what they output and fill gaps.

What should you check in a platform's output?

View a product page's source and the rendered markup, validate it, and compare the price, availability and identifiers with the page. Check what the feed contains for the same product. Confirm how the platform handles variants, sale prices, taxes and multiple currencies, because these are the most frequent sources of mismatch.

Do platform defaults include identifiers?

Often only if you fill the relevant fields. Missing GTINs are frequently a data entry gap and not a platform limitation.

When is custom work worth it?

When the platform cannot represent your catalog correctly, such as complex bundles, made-to-order items or multi-region pricing. Define requirements from Google's documentation first, then decide whether a plugin, a customization or a separate feed generator is best.

How do you avoid duplicate markup?

Themes, plugins and apps can each add Product markup, creating duplicates that conflict. Keep one source, disable the others and re-validate.

How do you manage multiple markets?

Set up separate feeds or settings by country and language where required, align currency and tax treatment with each market and keep localized pages consistent with their feeds. If you sell abroad, test each market separately, since a feed that is correct for one country can fail in another because of currency, tax display or shipping rules.

Who signs off on a new market?

The data owner and the person responsible for policies should both confirm that prices, taxes, delivery promises and return terms are correct for the new country before products are submitted, and that the localized pages show the same values as the feed. This sign-off takes little time and prevents a large batch of disapprovals on launch day, which is much more disruptive than checking beforehand.

Which checks apply to each market?

Price format, currency, shipping options, return terms and language of titles and descriptions.

How do accuracy and trust connect to visibility?

Shopping and search systems favor merchants whose data matches reality, because shoppers who find a different price or stock status lose trust. Accurate data is both a compliance matter and a conversion matter.

What do shoppers notice first?

Price, availability, delivery time, return terms and ratings. If any of these differs between the search result and the page, shoppers leave and may not return. Matching the information protects revenue as well as eligibility.

How do you check what shoppers see?

Search for your products on a phone, click through and compare the result with the page. Do it for sale items and out-of-stock items, which are the most likely to mismatch.

Which policies protect merchants?

Google's Merchant Center policies and the Search spam policies both penalize misleading data. Misrepresenting prices, availability or business identity can lead to disapproval or suspension. Keep your business name, contact details and legal information consistent between the website, Merchant Center and Organization markup.

What is the safest approach to claims?

Only state what you can deliver. Remove "free shipping" claims that carry hidden conditions, and show total costs clearly.

How do reviews fit in?

Genuine product and merchant reviews help shoppers and may be eligible for review features when marked up correctly and visible on the page. Never invent or purchase reviews, and keep markup aligned with Google's review snippet guidelines. Our article on reviews and Q&A for NLP covers how to collect and present them.

Can you display third-party review ratings?

Only if you follow the guidelines for the markup and the source of the reviews, and make sure that what is marked up is visible and accurate.

What quick wins can you make this week?

A few small actions often remove the biggest causes of disapprovals. They need little development time and give you a clearer view of the real state of your data.

Which five checks give the most value?

1. Compare ten random products across the page, markup and feed for price and availability.
2. Open Merchant Center diagnostics and fix the largest group of errors first.
3. Validate one page from each product template with the Rich Results Test.
4. Read your returns and shipping pages, then confirm that Merchant Center and Organization markup say the same thing.
5. Check that product images load for crawlers and that robots.txt does not block them.

What should you do with the results?

Log every finding with its cause, assign an owner and set a date. Group related findings so that one fix clears many items.

What can wait until later?

Large rebuilds, replatforming and advanced segmentation can wait until the foundation is accurate. Perfecting advanced setups on top of inconsistent data wastes effort.

How do you sell the work internally?

Point to the number of disapproved items, the products missing from results and the customer complaints about price differences. Concrete counts make the case for time from developers.

What should you document?

Write down where each field comes from, who owns it and how often it updates. Documentation turns a fragile setup into a process that survives staff changes and platform updates.

What does a clean-up look like in practice?

This is a fictional example that illustrates the method. Results are invented for demonstration.

What was the situation?

A fictional outdoor equipment store sees many feed items disapproved for price mismatches and notices that product snippets have disappeared for some pages.

What did the audit show?

The feed was generated nightly from the product database, but sale prices were applied on the site by a script that updated only the visible price. The markup showed the regular price. Stock status in the feed lagged behind the site, and several items lacked GTINs that existed in the supplier data.

What changed?

The store moved sale pricing into the database so that the page, markup and feed read the same field, increased feed refresh frequency, imported GTINs from supplier data, aligned the return and shipping information in Merchant Center, on the site and in Organization markup and added automated checks comparing sampled products across the three sources. They then monitored diagnostics weekly and noted that reprocessing and recrawling took time and that outcomes could not be guaranteed.

What was the key lesson?

One source of truth removed the root cause. Fixing individual items would have required repeating the work every week.

How does AH SEO Agency help ecommerce teams?

We audit product pages, markup, feeds and policy information, find the mismatches that cause disapprovals and lost features, and help your developers set up a single source of truth and monitoring. We use Google's documentation and say when something is unconfirmed.

Where do you start?

See our product page SEO service, our guide to AI shopping and feeds or request a free SEO audit.

Summary: how should you handle feeds and markup?

Treat the feed and the markup as two views of one catalog. Generate them from the same data, check them against the visible page, and monitor diagnostics regularly.

• Feeds send product data to Merchant Center, and markup describes products on pages.
• They should agree on price, availability, identifiers and policies.
• Organization markup can describe business-level return and shipping policies.
• Diagnose by comparing feed, page and markup for one item at a time.
• Automate generation and testing to prevent drift.