Merchant Center Feed Optimisation: The 12 Changes That Actually Move the Needle
Most feed optimisation guides hand you a spreadsheet with 47 attributes and tell you to fill them all in. They're not wrong, exactly — more complete feeds do perform better. But if you're looking at that list and wondering where to start, you're going to spin your wheels on low-impact fields while your competitors optimise the five things that actually matter for auction eligibility and CTR.
This post is different. These are the 12 changes that practitioners — people managing real budgets on real accounts — consistently find move the needle. Not in theory. In practice. Implement these before you worry about anything else.
1. Title Structure: Brand + Product Type + Key Attributes in the First 70 Characters
What to do: Restructure your product titles to follow this format in the first 70 characters: [Brand] [Product Type] [Key Attribute 1] [Key Attribute 2]. For example: "Samsonite Carry-On Luggage 20 Inch Hardshell Black" — 51 characters, every word earns its place.
Compare that to: "Hardshell carry on luggage in black colour - Samsonite brand - 20 inch size" — the brand and product type are buried after character 25. In most Shopping placements, Google truncates titles at approximately 70 characters. Everything after that is invisible to the user.
Why it matters: Google uses your title as a primary query-matching signal. Front-loading brand, product type, and key differentiating attributes means you're matching higher-intent queries from the start. Brand recognition in the first few words also improves CTR for branded searches — users scanning Shopping results find familiar names faster when they appear early.
Use Google's title recommendations in Merchant Center (Products → Title recommendations) as a diagnostic starting point. They flag when your titles are underperforming against Google's own benchmarks.
What to watch for: Don't sacrifice accuracy for character optimisation. Keyword-stuffed titles ("Yoga Mat Non-Slip Thick Eco Rubber Exercise Workout Gym Pilates Anti-Slip") trigger quality flags and erode user trust. Every word should accurately describe the product.
2. GTIN/MPN Presence for All Branded Products
What to do: Ensure every branded product in your feed has a valid GTIN (Global Trade Item Number — usually a barcode UPC, EAN, or ISBN). For products you manufacture without a GTIN, set identifier_exists = FALSE and populate the mpn (Manufacturer Part Number) field instead. Never leave both GTIN and identifier_exists blank for branded products.
Why it matters: GTINs unlock several auction advantages simultaneously. Google can match your product listing to other retailers selling the identical item, enabling accurate price benchmarking. This improves your competitive positioning signals and auction eligibility. Products with valid GTINs also get stronger consolidation — Google can show the right variant at the right time rather than guessing from your title.
Invalid or missing GTINs are one of the most common reasons for serving suppression. Google's policy requires GTINs for all products where one is assigned by the manufacturer. Omitting them doesn't just limit performance — it can result in disapproval.
What to watch for: GTIN validation errors. If you're pulling GTINs from your product database, verify them against GS1's GTIN database (gs1.org) before submitting. Fabricated or incorrect GTINs trigger policy violations that are harder to resolve than missing GTINs.
3. Keyword-Rich Titles Based on Actual Search Queries
What to do: Pull your Shopping search term report (In Google Ads: Insights → Search terms for PMax; in standard Shopping: Keywords → Search terms). Cross-reference with Google Search Console's Performance report for organic product queries. Identify high-volume, high-intent terms that your current titles aren't matching, and incorporate them as natural attribute additions.
"Yoga Mat" becomes "Non-Slip Yoga Mat 6mm Thick Eco-Friendly Natural Rubber" — not stuffed, just specific. Every added attribute is a real product feature that also happens to match search queries.
Why it matters: Google matches Shopping ads to queries primarily through your product title. Unlike Search ads where you write keyword-targeted copy, Shopping titles do the targeting work. More descriptive titles match more queries. This is especially critical in PMax, where Google's broad-match intent coverage depends heavily on what's in your feed.
What to watch for: Don't add attributes that aren't true of the product. Adding "eco-friendly" to a product that isn't eco-friendly to chase searches is a policy violation waiting to happen. Stick to genuine attributes — there are usually more than enough to work with.
4. High-Resolution Images (Minimum 800×800, Ideally 1200×1200+)
What to do: Replace all product images below 800×800 pixels with high-resolution versions. Target 1200×1200 or higher. Use white or neutral grey backgrounds for primary product images. Remove watermarks, promotional text overlays ("SALE 20% OFF"), and heavy decorative borders before submission.
Why it matters: Image quality has a direct, measurable effect on CTR in Shopping results — users are comparing multiple product images at a glance and clicking the ones that look cleanest and most professional. In PMax, the stakes are higher: Google uses your product images to auto-generate responsive display ads across Search, Display, YouTube, Gmail, and Discover. A low-resolution source image produces blurry ads at every placement size. That translates to lower engagement scores, which feeds back into auction competitiveness.
White backgrounds also enable Google's lifestyle image overlays — Google can add contextual backgrounds to your product image for certain placements, but only if the source image is clean.
What to watch for: Images with overlaid promotional text are a common disapproval trigger. If your e-commerce platform automatically adds watermarks or badges to product images, you'll need to configure a separate image source for your feed.
5. Granular product_type Values (More Specific Than Google Product Category)
What to do: Populate the product_type field with a specific, hierarchical classification that goes deeper than what Google's product taxonomy provides. Use your own category structure: "Women's Clothing > Shirts > Linen Button-Down Shirts" rather than just mirroring Google's Apparel & Accessories > Clothing > Tops.
Why it matters: Google's product categories are standardised — there are only so many of them, and they're necessarily broad. product_type is a free-form field you fully control, and Google uses it as an additional signal for query matching in PMax and for Shopping classification. The more granular your product_type, the more precisely Google can identify what your product is and which queries it should serve against.
This field is also invaluable for campaign segmentation. If you want to build separate Asset Groups or ad groups by product type, a well-structured product_type hierarchy makes that clean and maintainable.
What to watch for: Inconsistency. If some products have product_type set and others don't, your segmentation breaks down. Audit your feed for coverage and ensure your categorisation logic is applied consistently across all products.
6. Custom Labels for Segmentation (Not Decoration)
What to do: Use your five custom label slots (custom_label_0 through custom_label_4) for data that drives YOUR bidding and segmentation decisions — not data that's already available in other attributes. Recommended structure:
- Label 0: Performance tier (top-performer, mid-tier, low-performer, new-product)
- Label 1: Margin tier (high-margin, standard-margin, low-margin)
- Label 2: Price competitiveness (competitive, at-parity, premium)
- Label 3: Promotion status (on-sale, clearance, full-price)
- Label 4: Product lifecycle (hero, seasonal, end-of-life)
Why it matters: Custom labels are the mechanism for translating business intelligence into feed structure — and feed structure determines how much control you have over bidding by product segment. Without useful custom labels, you're bidding the same way on your 40% margin hero products and your 8% margin long-tail SKUs. That's money left on the table.
See our custom labels guide for detailed segmentation strategies.
What to watch for: The most common mistake is using custom labels to store data already in other fields — brand names, Google product categories, or condition. That wastes label slots on redundant data. Labels should encode information about your business context that Google's standard attributes can't capture.
7. Explicit condition Attribute for All Products
What to do: Set the condition attribute explicitly for every product. Valid values are new, refurbished, and used. Don't leave this field blank even for obviously new products.
Why it matters: When condition is missing, Google makes an assumption — and assumptions can result in misclassification. A product listed without a condition may be excluded from auctions it should be eligible for, or included in price benchmarking comparisons against used/refurbished items it shouldn't be compared to. For refurbished or used products, explicit condition is critical: these categories have specific auction eligibility rules, and incorrect classification leads to policy violations.
What to watch for: Bulk feeds from e-commerce platforms often omit condition entirely if the field isn't mapped. Check your feed export configuration and ensure condition: new is included in your default product template.
8. Real-Time availability Updates
What to do: Configure your feed to update availability status at least every 4-6 hours via scheduled fetch, or use the Content API for real-time updates. Set availability_date for preorder products to define when Google should activate serving. Remove discontinued products from your feed rather than leaving them in "out of stock" status indefinitely.
Why it matters: Serving out-of-stock products is one of the most damaging things you can do to your Shopping account health. Google tracks landing page experience signals — when users click through a Shopping ad and find the product unavailable, it registers as a negative experience signal. Enough of these signals at the account level damage your overall auction quality, affecting even the products that are in stock.
The practical impact: stale availability data doesn't just waste ad spend on non-converting clicks. It actively degrades the performance of your entire Shopping and PMax infrastructure.
What to watch for: High-velocity SKUs (flash sales, trending products, seasonal items) go out of stock faster than a 24-hour feed refresh cycle can track. For these categories, Content API integration or a supplemental availability feed with a very frequent fetch schedule is worth the implementation effort.
9. Rich, Long-Form Descriptions (Aim for 500+ Words)
What to do: Replace thin product descriptions with detailed, genuinely useful product information. Include material composition, dimensions, care instructions, compatibility information, use cases, size guide references, and any certification or standard compliance relevant to the product. Aim for 500-1000 words per product in your description field.
Why it matters: Standard Shopping ads display your title and image — descriptions aren't visible to users in Shopping results. So why does description length matter? Because Google's PMax algorithm uses your product description for query matching. This is where PMax gets its broad-match coverage — the algorithm reads your description to understand what queries your product should be relevant to, beyond what the title alone signals.
A thin description ("Blue cotton t-shirt. Available in sizes S-XL.") gives the algorithm almost nothing. A rich description covering fabric weight, weave type, care instructions, styling suggestions, ethical sourcing details, and size guide information gives Google substantially more signal for contextual matching across PMax's full inventory of placements.
What to watch for: Description content must accurately describe the product — it's still subject to Merchant Center's content policies. Don't add keyword-stuffed text or promotional claims ("Best yoga mat in Australia!") that violate editorial guidelines. Write for accuracy and completeness, not SEO manipulation.
10. sale_price + sale_price_effective_date for All Promotions
What to do: When running promotions, use the sale_price and sale_price_effective_date attributes rather than simply reducing your regular price. Set the sale_price_effective_date to the exact start and end timestamps of your promotion in ISO 8601 format (including timezone offset).
Why it matters: Using sale_price correctly triggers the price drop badge in Shopping results — strikethrough pricing that shows the original price next to the sale price. This visual treatment consistently improves CTR, with typical uplifts in the 15-30% range depending on category and discount depth. Users scanning Shopping results are highly responsive to visible price drops.
Setting exact effective dates means Google automatically activates the sale pricing display at the promotion start and reverts to regular pricing at the promotion end — no manual feed updates required mid-promotion.
What to watch for: If your sale price is less than the minimum discount threshold Google requires (typically at least a 5% reduction from the regular price), the price drop badge won't display even if the attributes are correctly formatted. Also watch for the price attribute — it must reflect your actual regular (non-sale) price for the strikethrough to display correctly.
11. Accurate shipping Attribute (Especially for Free Shipping)
What to do: Populate the shipping attribute for every product with accurate shipping costs, or configure shipping settings in Merchant Center (either at the account level or via feed override). If you offer free shipping above a cart value threshold, encode this as conditional free shipping in your Merchant Center settings rather than leaving the field blank.
Why it matters: Missing or inaccurate shipping data is a serving suppression trigger. Google requires shipping information for most product categories, and products without it may be disapproved or given limited serving status. Beyond eligibility, accurate shipping information affects user trust: if a user sees "Free Shipping" in your Shopping ad and then encounters a $15 shipping fee at checkout, conversion rate suffers and you've burned the ad spend.
For PMax specifically, accurate shipping contributes to the product-level quality signals that feed into auction eligibility calculations.
What to watch for: Flat-rate and tiered shipping structures require careful encoding in Merchant Center's shipping settings. Test your shipping configuration against representative product prices to verify it displays correctly before running live campaigns. Pay particular attention to weight-based shipping — if your feed doesn't include accurate product weights, weight-based shipping rules will produce incorrect quotes.
12. Supplemental Feed for Dynamic Custom Label Updates
What to do: Create a Google Sheets supplemental feed linked to your Merchant Center account. Use it exclusively for custom label updates. Structure the sheet with columns for id (matching your primary feed product IDs) and the custom label fields you need to update (custom_label_0 through custom_label_4). Set the fetch schedule to match how frequently your labels need updating.
Why it matters: Rebuilding your primary feed every time you need to update custom labels is slow, risky, and operationally expensive. A primary feed rebuild can take hours and introduces the risk of data errors that cause disapprovals. A supplemental feed targets only the fields you're updating — the primary feed data is untouched.
For keeping performance tier labels current (your Label 0), this is essential. As your account performance data updates weekly, you want to be able to reclassify products between tiers quickly. With a Google Sheets supplemental feed, that's a matter of updating the sheet and waiting for the next scheduled fetch — not a full feed engineering cycle.
What to watch for: Supplemental feed values override primary feed values for the fields they contain. If a product ID in your supplemental feed doesn't match a product ID in your primary feed, the supplemental row is ignored rather than creating a new product. Maintain a clean, current ID mapping to avoid ghost rows in your supplemental feed.
Start Here
These 12 changes represent the highest-leverage feed optimisation work available to most Shopping and PMax advertisers. They're not exhaustive — a thorough feed audit covers far more attributes — but they consistently move the needle in real accounts. Implement them in order: title structure and GTINs first, descriptions and supplemental feed last. Measure the impact of each change before moving to the next, and you'll have clear attribution data for what's actually driving your improvements.