Shopify Product Feeds for Meta and Google: Getting the Catalogue Right

Your Meta and Google campaigns are not doing what the spend suggests they should. The media buyer says it is creative, or audiences, or the algorithm needs more time. Sometimes that is true. More often the problem sits underneath the campaign in the Shopify product feed: the data Google and Meta use to decide what to show, to whom, and whether to show it at all.

This is not a media buying article. We do not run paid campaigns at Uncover. What we do is build and maintain the store data those campaigns depend on, and feed quality is a build and data job. If the catalogue is wrong, no bid strategy fixes it. Feed quality sets the ceiling on paid performance.

How product data gets from Shopify to Google and Meta

Most stores use the two native sales channels: Google & YouTube, and Facebook & Instagram. Both read product data from Shopify and push it to Google Merchant Center and Meta Commerce Manager. The Google channel re-syncs a product whenever it changes, and refreshes everything inside Google's 30-day expiry window so listings do not silently drop out. Meta works the same way, with the catalogue tied to your Pixel and Conversions API data.

The important point is that the channels send what Shopify has. They do not invent a brand field, tidy a title, or guess a GTIN. Third-party feed tools (Feedonomics, DataFeedWatch, Simprosys and others) give you more control over mapping and rules, but they still start from the same source data. Fix the data in Shopify and every destination benefits at once.

Required attributes, and the ones that actually decide performance

Google's product data specification requires id, title, description, link, image link, price and availability for every product, plus brand for almost everything. Meta's list is close to identical. Getting approved is the easy part. Performing well depends on the recommended attributes that most Shopify stores leave half empty.

GTIN and MPN. Google checks GTINs against the GS1 database to place your product alongside identical listings. When Google made GTINs mandatory for branded goods in 2016, it reported that products with a valid GTIN saw up to 40% more impressions and 20% more conversions. In Shopify this lives in the barcode field. If a product genuinely has no GTIN (own-brand, handmade), set identifier exists to no and supply brand plus MPN. Never reuse one GTIN across variants, and never invent one. Google catches both.Google product category and product type. Category tells Google what the item is, for policy and matching. Product type is your own taxonomy and is one of the most useful fields for structuring campaigns.Colour, size, gender, age group and material. Required for apparel and useful everywhere. In Shopify these now map from category metafields, which unlock when you assign a standard product category to the product.Sale price and sale price effective dates. Run a promotion through Shopify discounts without these and the feed price disagrees with the page price. That is a disapproval waiting to happen.

Titles and descriptions written for the feed, not lifted from the page

A product page title is written for someone already on your site. A feed title has to win a search it never sees. Google accepts up to 150 characters in a title but displays roughly the first 70, so brand, product type and the defining attributes need to be at the front. For apparel that usually means brand, gender, product type, colour, size. For hard goods: brand, product type, key specification, model number.

Most Shopify stores send the page title unchanged. "The Islay Jacket" tells Google almost nothing. "Brand Name Women's Waxed Cotton Jacket, Olive, UK 12" tells it everything it needs to match the query. Descriptions follow the same logic: plain, complete, specific, no HTML, no promotional copy. If you want the on-site title to stay short and brand-led, store a feed title in a metafield and map that instead. You do not have to choose between the two.

Variants, colour and images

Each variant should be its own item in the feed, grouped under the parent by item group id. Shopify handles the grouping, but it only sends what each variant actually has. A variant with no image of its own inherits the parent image, so someone searching for the navy version is shown the black one. Assign variant images, and keep colour values consistent across the catalogue (not "Navy" on one product and "Dark Blue/Navy" on the next).

Image rules are stricter than most stores assume: no promotional overlays, no watermarks, no placeholder images, and a clear product shot as the primary image. Google's minimum is 100 by 100 pixels for most categories and 250 by 250 for apparel, with 800 by 800 or larger recommended. Lifestyle imagery often performs better on Meta, but Google needs a clean primary image to approve and match the product reliably.

Custom labels: segment by margin, season and stock position

Custom labels 0 to 4 are five free-text fields that exist purely so campaigns can be split. In Shopify they are stored as product metafields and synced through the channel. Used well, this is where a feed starts to pay for itself:

Margin band (high, mid, low) so whoever runs the ads can bid harder on products that can afford it.Stock position (overstocked, core, low) so you stop paying to send traffic to items about to sell out.Season or launch flag so new ranges get budget before they have any performance history.Price tier or best seller flag for campaigns built around proven products.

For a small catalogue you can populate these by hand. For a large one, a Shopify Flow automation keyed off inventory and cost fields keeps them current without anyone having to remember.

Feed disapprovals: the most common source of silent lost revenue

A disapproved product does not fail loudly. It just stops appearing. It is common to find a meaningful slice of a catalogue disapproved for months, including products the campaigns were built around, because the reports still show spend and conversions. Just fewer than there should be.

Price or availability mismatch. The feed says one thing and Google's crawl of the page says another. Usually caused by regional pricing or currency conversion through Shopify Markets, promotions set without sale price fields, or a change that has not synced yet.Invalid or duplicated GTIN. Wrong digits, one barcode copied across every variant, or a GTIN that belongs to a different product entirely.Missing required attributes. Brand, colour or size absent on an apparel product. Blank category metafields are the usual culprit.Image violations. Text on the image, overlays, too small, or a generic placeholder.Policy conflicts. Restricted categories, claims in descriptions that Google treats as misleading, or landing pages that behave differently from the feed.

Clearing them is a weekly job, not a one-off. Merchant Center's Needs attention view and Meta's catalogue diagnostics both list the reason per item. Work from the highest-revenue products down, fix the data in Shopify rather than overriding it in the platform, and resubmit. If the same error keeps returning, it is a sync or template problem rather than a product problem, and that needs a developer.

Metafields take the feed beyond Shopify's defaults

Shopify's standard product fields were never designed to carry everything a feed wants. Metafields fill the gap: feed-specific titles, material, pattern, energy efficiency class, unit pricing, bundle flags and the custom labels above. The native Google channel maps a fixed set of them. Feed apps will map almost anything. Either way, define the fields once, populate them properly, and treat them as part of the product record rather than as marketing's problem.

One piece of work, three payoffs

The reason this matters at a store level is that the structured, complete product data that makes a feed perform is the same data Google's organic results read, and the same data AI shopping assistants read. A product with a correct GTIN, a specific title and a full set of consistent attributes is easier for Merchant Center to approve, easier for search to rank, and easier for an assistant to recommend. We covered the AI side in how to structure product data for AI shopping agents. It is the same work.

It also feeds back into the store itself. Clean attributes power filters, comparison tables and the specifics on the product page that actually convert. If your product pages are missing the details buyers need, the feed is almost certainly missing them too. Most Shopify brands do not have a traffic problem. They have a conversion problem, and paid traffic landing on thin product data is part of it.

Underperforming spend is usually a data problem before it is a creative problem.

What to do this week

Pull the disapproval list from Merchant Center and Commerce Manager and add up the revenue tied to those products. Check whether your barcodes are real GTINs. Look at the feed titles for your top 20 products and ask whether a stranger could tell what they are from the first 70 characters. That audit takes an afternoon and tells you whether you have a data problem.

If you do, it is a build job. We handle feed and catalogue work as part of Shopify store improvements and upgrades, and on an ongoing basis through our Conversion Growth Retainer, where product data sits alongside the on-site conversion rate optimisation it supports. We are a senior team based in Glasgow, working with brands across the UK and beyond, and we have been doing this for over 12 years. If paid results have plateaued and nobody has looked at the catalogue, get in touch and we will tell you honestly whether the data is the problem.