We audited a mid-size home goods retailer's Shopping campaigns last spring expecting to find a bidding problem. Their account manager had been tweaking target ROAS settings for months with barely any movement. The actual problem was sitting in the product feed the entire time: vague titles, missing GTINs on a third of the catalog, and product types that didn't match how Google's algorithm actually categorizes search intent. Google Shopping optimization gets treated as a bid-and-budget exercise by a lot of advertisers, when the feed itself is usually where the real ceiling on performance gets set. We didn't touch a single bid for the first three weeks of that engagement. ROAS still climbed 68% just from fixing what was in the feed.
Key Takeaways
- Missing or generic GTINs on just 30% of a catalog reduced overall Shopping impression share by roughly 40% for the retailer we audited, since Google heavily favors feed data it can verify against manufacturer records.
- Titles built around actual search query language, not internal SKU naming conventions, drove a measurable share of the eventual ROAS lift, independent of any bid changes.
- Product type and Google product category mismatches routinely send traffic to the wrong shopping tab or suppress listings entirely from relevant searches.
- High-resolution images with a clean, consistent background outperformed lifestyle imagery on click-through rate for this catalog, reversing what the brand's creative team originally assumed would work best.
- Feed health should be audited on a monthly cycle at minimum, since price mismatches, stockouts, and attribute drift accumulate quietly and quietly tank performance.
The Data Behind the Feed Problem
A Google Shopping optimization strategy lives or dies on data quality long before it ever touches a bidding algorithm. Google's Shopping algorithm relies almost entirely on structured feed data to decide which queries a product matches, and unlike search ads, there's no keyword targeting to fall back on when the data is thin. If the feed doesn't spell out what a product is, who it's for, and what makes it distinct from a near-identical competitor listing, the algorithm fills in the gaps poorly.
In the audit, we pulled impression share data segmented by product and cross-referenced it against feed completeness scores. Products with fully populated GTINs, detailed product types, and custom labels captured impression share at nearly double the rate of products missing that data, even when priced comparably against competitors. That gap alone explained a huge chunk of the account's underperformance, and it had nothing to do with how aggressively the account was bidding.
Custom labels turned out to be doing more work than the account team expected, too. The feed had five custom label slots sitting completely unused, which meant there was no way to segment campaigns by margin, seasonality, or bestseller status inside the bidding structure. We rebuilt the labeling scheme around margin tier and inventory velocity, which let the campaign structure prioritize spend toward products that were both profitable and actually in stock, instead of spreading budget evenly across a catalog where a third of the items were backordered.
Finding 1: Title Structure Is Doing More Work Than Most Advertisers Realize
If you're wondering how to do Google Shopping optimization on your own feed, start with titles before anything else. Google's algorithm weighs the first several words of a product title heavily when matching to search queries, which means titles built around internal naming conventions (brand-first, model-number-second) routinely underperform titles structured around how customers actually search: attribute, brand, then product type, in an order that mirrors real query patterns.
We rebuilt titles across the home goods catalog using a consistent template: primary attribute (material, size, or color), brand, product type, and a distinguishing detail, replacing titles that had previously led with an internal SKU code nobody outside the warehouse ever typed into a search bar. Click-through rate on the revised listings climbed noticeably within the first two weeks, well before any bid adjustments were made.
Finding 2: A Product Feed Setup That Actually Converts Needs Complete, Verified Identifiers
A product feed setup that actually converts treats GTINs, MPNs, and brand fields as non-negotiable, not optional extras. Roughly a third of the retailer's catalog was missing GTINs entirely, mostly on private-label items where the team assumed the field didn't apply. Google's Shopping algorithm treats products without verified identifiers as inherently less trustworthy matches, which suppresses both impression share and, in some categories, eligibility for certain placements altogether.
For private-label products genuinely without a manufacturer GTIN, marking the "identifier_exists: false" field correctly (instead of leaving it blank or filling it with a placeholder) resolved most of the suppression. That single fix alone accounted for a meaningful share of the impression share recovery, because Google stopped penalizing the listings for data it had previously flagged as incomplete or suspicious.
Finding 3: Image Quality and Background Consistency Change Click Behavior More Than Expected
The brand's creative team had pushed for lifestyle photography across the catalog, assuming it would perform better than plain product shots. Testing showed the opposite for this specific catalog: clean, high-resolution images on a neutral background outperformed lifestyle shots on click-through rate by a wide margin, likely because shopping search intent tends to be comparison-driven, and a cluttered lifestyle image makes it harder to quickly assess size, color, and shape against competing listings in the grid.
This isn't universal across every vertical, apparel often does better with some lifestyle context, but it's a reminder that assumptions about creative performance need to be tested against the specific product category and search behavior, not applied as a blanket rule.
Price accuracy belongs in this same conversation even though it's not technically a creative element. Google disapproves or suppresses listings the moment a feed price drifts too far from the price shown on the landing page, and for a catalog with frequent promotions, that mismatch happens more often than teams realize. We set up a daily price-check script for the home goods client after finding that roughly 4% of active listings were mismatched on any given day, mostly from sales that ended on the site before the feed update caught up. That 4% doesn't sound like much until you realize it was concentrated in the catalog's higher-margin bestsellers, the exact products the account most needed showing up in search.
What This Means for Channel Strategy
The broader lesson here isn't really about Shopping ads specifically, it's about where teams choose to spend their optimization time. Bid strategy adjustments feel like the "real" lever because they're the most visible dial in the account. Feed quality work is slower, less glamorous, and often owned by a different team (product or ecommerce ops) than the one running the ads, which means it gets neglected even though it sets the ceiling everything else operates under.
Best practices for Google Shopping optimization increasingly point toward treating the feed as a living asset that needs regular maintenance, not a one-time setup task completed at launch and forgotten. Price changes, stockouts, and seasonal attribute updates all degrade feed quality quietly over time if nobody's checking.
Ownership matters as much as process here. Feed quality tends to fall through the cracks precisely because it sits between two teams: the ecommerce or product team that owns the underlying catalog data, and the ads team that owns the campaigns pulling from it. Assigning one person clear responsibility for feed health, even if it's just a monthly checklist review, closes that gap faster than any tool or automated feed rule on its own.
FAQ
Q: How often should a Shopping product feed be audited?
A: Monthly at a minimum for active catalogs, and weekly for larger catalogs with frequent price or inventory changes. Feed drift happens quietly and compounds fast if it goes unchecked.
Q: Does feed optimization matter more than bidding strategy?
A: For most underperforming accounts we've audited, yes, at least as a starting point. A well-bid campaign built on a weak feed still underperforms a modestly-bid campaign built on a clean, complete one.
Q: Should every product have a GTIN?
A: Every product with a genuine manufacturer identifier should include it. For private-label or custom products without one, correctly flagging "identifier_exists: false" prevents Google from penalizing the listing for missing data it never expected to find.
Q: Is lifestyle photography always worse than plain product images for Shopping ads?
A: No, it depends heavily on category. Comparison-heavy categories like home goods or electronics tend to favor clean product shots, while apparel and lifestyle products often benefit from some contextual imagery. Test rather than assume.
If your Shopping campaigns have stalled despite bid adjustments, the feed is worth a hard look before anything else. Our Google Ads management team runs full feed audits as a standard part of Shopping engagements, and pairs that work with broader ecommerce marketing strategy to make sure the fixes carry through to the rest of the funnel. Reach out through our contact page and we'll take a look at what's actually happening in your feed.
