The Reach Bureau

Google Shopping Feed Optimization: The Product Data That Decides Visibility

Google Shopping feed optimization for ecommerce stores

Your product feed is the version of your catalogue that Google actually reads, and for most stores nobody has ever looked at it.

It gets generated by a plugin from whatever fields happen to exist, uploaded, and forgotten. Products get disapproved silently. Titles arrive truncated. Attributes that decide whether you appear for a filtered search sit empty. None of this shows up on your website, so nothing prompts anyone to check.

The feed is also increasingly the same data that AI shopping surfaces consume. Getting it right has stopped being a paid-media chore and become part of how a store is understood.

The attributes that actually matter

Not all fields carry equal weight. These are the ones that change outcomes.

title — the single highest-impact field. It is matched against queries, and it is truncated in display, so the order of words decides what a shopper sees. More on this below.

description — used for matching, not just display. The manufacturer's paragraph works against you here for the same reason it does on your product page.

google_product_category — Google's own taxonomy. Getting this wrong puts you in the wrong competitive set. Use the most specific applicable value, not the closest broad one.

product_type — your taxonomy. Unlike the above, this is yours to define, and it is what campaign structure hangs off.

gtin / mpn / brand — identifiers. Missing GTINs on branded products is a common cause of poor matching, because Google cannot connect your listing to the product record everyone else is bidding on.

availability and price — must match the landing page exactly. Mismatches cause disapproval, and they are the most frequent avoidable error.

Attribute fieldscolor, size, material, age_group, gender. These power filtered browsing. An empty size field removes you from every size-filtered search.

The feed attributes ranked by impact — and the ones stores most often leave empty

Fixing titles at scale

Titles carry the most weight and are the most consistently wrong, because they are usually just the product name from your CMS.

The pattern that works puts the words a shopper searches first, because titles truncate:

Brand + Product + Key Attribute + Size/Variant

"Merrell Moab 3 Waterproof Hiking Shoe — Men's 4E Wide" beats "Moab 3" and beats "The Best Hiking Shoe for Every Adventure — Merrell."

Rules worth applying across the catalogue:

  • Brand first for branded goods; category first for unbranded
  • The distinguishing attribute early — waterproof, wide fit, organic, refurbished
  • No promotional language: "sale", "free shipping", "best" get products disapproved
  • No ALL CAPS, no decorative characters
  • Keep the important part inside roughly the first 70 characters

Do it with rules, not by hand. Most feed tools let you build titles from fields. A rule that assembles brand + title + colour + size across four thousand products is an afternoon; editing four thousand titles is not.

The errors that quietly suppress products

Disapprovals are visible in Merchant Center. The quieter problems are the ones worth hunting.

ProblemSymptomFix
Price mismatchDisapproval, or intermittentFeed price must match landing page including currency and tax display
Missing GTIN on branded goodsPoor matching, low impressionsAdd identifiers; use identifier_exists: no only when genuinely absent
Wrong google_product_categoryAppearing in the wrong comparisonsUse the most specific applicable node
Empty attribute fieldsAbsent from filtered searchesPopulate color, size, material at minimum
Landing page out of stockDisapprovalFeed availability must reflect reality at crawl time
Truncated titleLow click-throughRestructure so the identifying words come first
Image with overlay textDisapprovalClean product images only, no badges or watermarks

The audit worth running monthly: export the feed, count how many products have an empty value in each attribute column, and sort by which gaps affect the most products. That single view usually reveals more than the diagnostics dashboard.

Where the feed meets AI shopping

Assistants answering shopping questions need structured facts, and product feed data is the cleanest structured description of a catalogue that exists.

The overlap is direct: the same completeness that makes a feed perform in Shopping — populated attributes, accurate availability, specific titles, real identifiers — is what lets an assistant match your product against a constrained query like "waterproof hiking boot, wide fit, under $150."

Two practical implications:

  • Consistency across surfaces matters. If your feed says one price and your product page's structured data says another, you have given contradictory inputs, and contradictions get dropped rather than reconciled.
  • The same discipline serves both. There is no separate AI feed. Per Google's Product structured data documentation, your on-page markup should carry price and availability too — and it should match the feed exactly.
How feed data, on-page structured data and AI shopping surfaces need to agree

The checklist

  • Feed exported and reviewed manually, not just via the dashboard
  • Empty-value count per attribute column measured
  • title built by rule: brand, product, key attribute, variant
  • Identifying words inside the first ~70 characters
  • No promotional language or ALL CAPS in titles
  • description original, not the manufacturer's paragraph
  • google_product_category set to the most specific applicable node
  • product_type reflects your own taxonomy for campaign structure
  • gtin, mpn and brand populated on branded products
  • identifier_exists: no used only where genuinely absent
  • price and availability match the landing page exactly
  • color, size and material populated
  • Images clean — no overlay text, badges or watermarks
  • On-page Product + Offer markup matches feed values
  • Monthly export-and-audit scheduled

Sources

Frequently Asked Questions

Improving the product data you send to Merchant Center so products match more queries, appear in more filtered searches, and stop being suppressed by errors. The highest-impact fields are title, category, identifiers and the attribute fields most stores leave empty.
Most commonly a price or availability mismatch between the feed and the landing page, missing identifiers on branded goods, or an incorrect product category putting you in the wrong competitive set. Disapprovals appear in Merchant Center; the quieter suppression from empty attributes does not.
Brand, product, key distinguishing attribute, then size or variant — with the identifying words inside roughly the first 70 characters, because titles truncate. Build them by rule from existing fields rather than editing them individually.
It is the cleanest structured description of your catalogue that exists, and the same completeness that performs in Shopping — populated attributes, accurate availability, specific titles, real identifiers — is what lets an assistant match a constrained query.
Export it and count empty values per attribute column, then sort by which gaps affect the most products. That view usually surfaces more than the diagnostics dashboard, because empty fields suppress visibility without generating an error.
Yes, exactly — price, availability and identifiers. Mismatches cause disapproval, and where your on-page structured data disagrees with your feed you have given contradictory inputs that tend to be dropped rather than reconciled.

Want this run against your store? Book a call with The Reach Bureau.

Share with AI

One-minute takeaway Summarize Explain like I'm a kid

Share this article

LinkedIn X Facebook Pinterest Email

Related articles

View all articles

Ready to scale your e-commerce?

Let's discuss your project and how we can help you achieve your growth goals.

Book a discovery call
Book a call with me, here is my schedule →