How to build a Google Shopping feed for WooCommerce

Merchant Center rejections are rarely mysterious. Google publishes exactly what it wants, checks it mechanically, and disapproves anything that does not match. Almost every rejection falls into one of four buckets: a missing required attribute, a missing product identifier, a mismatch between the feed and the landing page, or a policy issue with the product itself.

Here is how to build a WooCommerce feed that clears all four.

The required attributes

For a standard retail product sold online:

Apparel and accessories add color, size, age_group, gender and size_type. If you sell clothing and skipped these, that is your disapproval.

Product identifiers, the number one rejection cause

Google wants to know which real-world product each listing refers to, so it can group offers and show comparison results. It infers this from unique product identifiers.

The rule:

• If the product has a manufacturer-assigned GTIN (a UPC, EAN or ISBN), you must supply it. Omitting a GTIN that exists is a disapproval, not a warning.
• If you manufacture the product yourself and there is genuinely no GTIN, supply brand plus mpn and set identifier_exists to no.
• Do not invent GTINs. Google validates the check digit, and a wrong one is worse than an absent one.

In WooCommerce, GTIN lives in the barcode field on modern versions, or in a custom field on older ones. If your data came from a migration, check it survived: barcode is one of the fields most commonly dropped in a CSV round trip.

Variants

Anything a customer chooses between needs its own feed entry. A shirt in four sizes and three colours is twelve entries, not one.

Each entry needs:

• Its own unique id, typically the variation SKU.
• A shared item_group_id across the whole group, typically the parent SKU.
• Its own link, ideally with the variation pre-selected: ?attribute_pa_size=large.
• Its own price, availability and image_link where these differ.
• The distinguishing attributes filled in: size, color, material, pattern.

Submitting only parent products is a common shortcut that costs you impressions, because Google cannot match a size-specific query to a listing that has no size.

Price and availability mismatch

Google crawls your landing pages and compares what it finds against what you submitted. A discrepancy triggers a warning, then a disapproval, and persistent mismatches can suspend the account.

The usual causes are all timing problems:

• The feed refreshes daily, prices change hourly.
• A sale started and the feed still carries the old price.
• Stock ran out and the feed still says in_stock.
• Currency or tax display differs between feed and page.

This is why refresh frequency is not a nice-to-have. If your prices or stock move, a daily feed is not enough, and a feed you upload by hand is a suspension waiting for a busy week.

Use XML, not a spreadsheet

Merchant Center accepts RSS 2.0 and Atom 1.0 XML from a URL it fetches on a schedule. This is better than uploading a file for reasons that have nothing to do with the format itself:

• Google pulls on a schedule, so there is nothing to remember.
• XML has unambiguous encoding and no delimiter problems, which removes an entire class of CSV failure.
• Nested attributes such as shipping and tax are expressible without inventing column conventions.

A minimal item looks like this:

<item>
  <g:id>NB-TEE-M-BLK</g:id>
  <g:item_group_id>NB-TEE</g:item_group_id>
  <title>Northbound Premium Cotton Tee, Black, Medium</title>
  <description>Heavyweight 240gsm combed cotton tee...</description>
  <link>https://example.com/product/cotton-tee?attribute_pa_size=medium</link>
  <g:image_link>https://example.com/img/tee-black.jpg</g:image_link>
  <g:availability>in_stock</g:availability>
  <g:price>24.99 USD</g:price>
  <g:brand>Northbound</g:brand>
  <g:gtin>0123456789012</g:gtin>
  <g:condition>new</g:condition>
  <g:size>Medium</g:size>
  <g:color>Black</g:color>
</item>

Generating the feed

There are two honest ways to produce a Merchant Center feed from WooCommerce, and both start from the same precondition: the products have to already exist in your store.

A feed plugin

If every product already lives in your WooCommerce database and you only need to reshape it into Google's attribute names, a feed plugin does that job. It reads your existing products and generates an endpoint Merchant Center can fetch on a schedule. For most stores this is the right answer and the end of the decision.

The limits worth knowing before you buy one: it publishes only what is already in your store, it cannot fix missing GTINs or thin descriptions, and refresh frequency is bounded by WP-Cron, which on a low-traffic site fires late or not at all. That last one is how a "daily" feed quietly becomes a weekly one and your account picks up mismatch warnings.

Build it yourself

A WordPress endpoint that queries products and echoes XML. Total control, and you own the maintenance forever. Worth it only if your requirements are genuinely unusual.

The step before the feed: getting the products in

Neither route helps if the catalogue is not yours yet. A supplier range, a dropship list, a store you are mirroring: a feed plugin cannot publish products that are not in your WordPress database, so there is nothing to reshape. People hit this constantly and mistake it for a feed problem when it is a catalogue problem.

That earlier step is what Scrapify covers. Scrape any WooCommerce or Shopify store, then our free WordPress plugin imports the catalogue into WooCommerce for you: products, variations, images, categories and prices, matched on SKU so nothing duplicates. A live feed on your account keeps it refreshing hourly, so supplier price and stock changes land in your store rather than drifting out of date, which is also what keeps your Shopping feed matching your landing pages.

Getting approved faster

1. Submit a 10 product test feed first. Fix what it flags before submitting 5,000 and receiving 5,000 errors.
2. Complete the account setup. Verified and claimed website, shipping settings, returns policy, tax. Google will not approve products in an incomplete account regardless of feed quality.
3. Match the landing page exactly. Same price, same currency, same availability, same product.
4. Fix the Diagnostics tab weekly. Warnings become disapprovals if ignored.
5. Write titles for search, not for shelves. "Northbound Premium Cotton Tee, Black, Medium" beats "Premium Tee" because it contains the words people type.

One feed, several uses

Once the feed exists it is worth more than Merchant Center alone. The same URL typically feeds Facebook and Instagram catalogues, Pinterest product pins, Bing Merchant Center, affiliate networks, and comparison shopping engines.

Each platform wants slightly different attribute names, but the underlying data is identical, and getting it right once pays out across every channel.

Frequently asked questions