Google Merchant Center Feed Optimization: A Practical Guide

Google Merchant Center feed optimization in 2026: what has changed

Google Merchant Center feed optimization has always been about data quality, but the stakes shifted in 2026. Google now uses feed data not only for Shopping ads and free listings but also for AI-powered surfaces like Gemini recommendations, AI Mode, and the new Business Agent. In January 2026, Google announced dozens of new data attributes designed for conversational commerce, including product FAQs, compatibility data, and substitute mappings. The message is clear: Merchant Center is evolving from a feed upload tool into a product intelligence hub.

For ecommerce teams managing hundreds or thousands of SKUs, this means feed optimization is no longer a one-time setup task. It is a continuous operations discipline that touches titles, identifiers, images, pricing, availability, taxonomy, and now conversational attributes. If your feed workflow still relies on manual CSV edits and reactive troubleshooting, the gap between your catalog and what Google expects will only widen.

This guide covers the practical steps that matter most. If you want a broader view of feed operations across channels, start with our product feed management overview first.

Title and taxonomy optimization for search relevance

Product titles are the single highest-leverage attribute in your feed. Google weighs them heavily for query matching in both Shopping ads and Performance Max campaigns. A title that satisfies the required character limit but ignores search intent is leaving impressions on the table.

Build a title template for each major product category. The most effective structure for most verticals follows this pattern:

Brand + Product Type + Key Differentiator + Variant Details

For example:

• Weak: "Running Shoes"
• Stronger: "Nike Pegasus 41 Running Shoes Men Black Size 10"

Guidelines that improve title performance:

• Front-load high-intent terms. Google truncates titles at different lengths depending on placement. The first 70 characters carry the most weight.
• Remove promotional language. Phrases like "Best price" or "Free shipping" violate Merchant Center policies and risk disapproval.
• Keep variant attributes consistent. Size, color, and material should appear in the same order across sibling SKUs so your catalog looks structured, not chaotic.
• Avoid keyword stuffing. Readability still affects click-through rate, and Google's algorithms detect unnatural repetition.

Pair title work with clean taxonomy. Set google_product_category to the most specific valid path in the Google taxonomy tree. Use product_type as your internal merchandising hierarchy for bidding segmentation and reporting. Version your category mappings so that taxonomy updates do not break historical performance data. For a detailed breakdown of title structures per category, see our title template guide.

Scaling title and taxonomy logic across a multi-source catalog is where teams hit a wall. Lasso solves this by applying rule-based title templates and taxonomy mappings during feed preparation, so merchandising teams stop rebuilding the same formulas in spreadsheets every export cycle.

GTIN validation and identifier hygiene

GTIN errors remain one of the most costly classes of feed problems. They affect both product eligibility and matching quality, because Google uses GTINs to link your listing to a canonical product in its catalog. Bad identifiers mean your products either get disapproved or compete poorly against merchants with correct data.

For every SKU, answer one question: does a valid manufacturer-assigned GTIN exist?

If yes:

• Submit the GTIN in the correct format (EAN-13, UPC-A, ISBN, or ITF-14 depending on region and product type).
• Validate checksum digits before export. A single transposed digit will cause rejection.
• Ensure the brand field matches the GTIN's registered manufacturer exactly.

If no:

• Set identifier_exists to false explicitly.
• Provide strong fallback identifiers: brand plus MPN at minimum.
• Document the no-GTIN status for audit purposes and category owner approval.

Operational controls that prevent recurring identifier errors:

• Run automated checksum and length validation in your feed pipeline, not inside Merchant Center after submission.
• Detect and block duplicate GTINs assigned to unrelated products.
• Flag placeholder values like all-zero strings, repeated digits, or obviously fabricated codes.
• Monitor Merchant Center diagnostics for the "incorrect GTIN" warning cluster weekly.

We covered detailed no-GTIN workflows and marketplace-specific handling in our missing EAN/GTIN guide. For teams managing supplier data at scale, a data cleansing pipeline that catches these problems before feed generation is significantly more efficient than fixing them after disapprovals appear.

Image quality: compliance first, conversion second

Images serve two purposes in Google Shopping: they must pass policy checks and they must attract clicks. Focus on compliance first, because a disapproved product generates zero impressions regardless of how compelling the photo is.

Google's current image requirements:

• Minimum 100x100 px for non-apparel, 250x250 px for apparel. Recommended: at least 1500x1500 px for high-quality Shopping placements.
• No watermarks, promotional overlays, badges, or embedded pricing text.
• The primary image must show the actual product on a clean background, not a lifestyle scene (lifestyle images belong in additional_image_link).
• Variant images must match the specific variant being sold. A red shoe listing with a blue shoe photo triggers misrepresentation flags.

Build two safeguards into your workflow:

1. Automated image validation that checks URL accessibility, file format, minimum resolution, and detects text overlays before export.
2. Manual spot review for your top-revenue SKUs before major seasonal campaigns, since these products carry the highest disapproval cost.

Image drift is a common problem in larger catalogs. A merchandising team updates product photos on the storefront, but the feed export cache still references stale image URLs. Treat image publication and feed refresh as a single deployment event to prevent this.

Price and availability synchronization

Price and availability mismatches are the fastest path to bulk disapprovals. These errors rarely come from bad source data. They come from timing gaps between systems: your storefront publishes a price change, structured data on the landing page updates on the next cache clear, and the feed export job runs on its own schedule. Google crawls the landing page, sees a different price than what the feed declares, and flags the mismatch.

A reliable synchronization model has four components:

1. Single source of truth. One system owns the current sellable price and stock status. Every downstream consumer, including your feed, reads from this source.
2. Deterministic export schedule. Know exactly when your feed updates reach Merchant Center. If you run hourly scheduled fetches, your maximum price drift window is one hour plus crawl delay.
3. Landing page parity checks. After each feed export, verify that a sample of product landing pages show prices and availability that match the feed. Automate this check for high-revenue products.
4. Threshold alerting. Set up monitoring that triggers when mismatch rates exceed a percentage you define. Do not wait for the Merchant Center diagnostics email.

Also review your Merchant Center automatic item updates setting. Google offers this as a safety net: if it detects a mismatch, it can pull the landing page value. This is useful as a temporary buffer, but relying on it permanently masks data quality problems and gives Google's crawler the final word on your pricing data.

For teams that syndicate to multiple channels, centralizing feed logic in Lasso makes it possible to enforce shared validation rules and normalize attribute values across platforms without manually reconciling exports. This is especially important when promotional pricing or stock allocation differs by channel.

Preparing for conversational commerce attributes

The biggest structural change in 2026 is Google's push toward agentic commerce. With the Universal Commerce Protocol and new AI surfaces, Merchant Center is adding attributes that go beyond traditional keyword-and-category feed logic.

New attribute categories include:

• Product FAQs: structured question-and-answer pairs that help AI agents respond to shopper queries about your products. Aim for 5 to 10 pairs per product, covering questions that other attributes do not answer.
• Compatibility data: which accessories, parts, or consumables work with a product. Critical for electronics, automotive parts, and home improvement categories.
• Substitute products: alternative items you stock that serve a similar purpose. This helps Google suggest your catalog even when the exact queried product is unavailable.

These attributes are rolling out to retailers in phases. Even if your account does not have access yet, start building the data now. FAQ content, compatibility mappings, and substitute logic take time to assemble for large catalogs, and early adopters will have a structural advantage when these signals start influencing ranking.

Think of this as an extension of your product data quality checklist: the set of required fields is growing, and the definition of a "complete" product record now includes conversational signals.

A weekly feed health workflow

Treating feed optimization as a periodic project instead of an ongoing operation is the most common failure mode. Build a weekly cadence that catches problems early and compounds small improvements over time.

Weekly feed health checklist:

1. Export and classify diagnostics. Pull Merchant Center issues and group them by root cause, not by individual SKU. Most accounts see the same five to ten error patterns repeatedly.
2. Prioritize by revenue impact. Fix the root causes that affect the most revenue first, not the ones that affect the most SKUs. A disapproval on your top 50 products matters more than a warning on 500 long-tail items.
3. Patch source rules, not channel UI. When you find a title pattern that triggers warnings or an identifier class that causes disapprovals, fix the rule in your feed pipeline. Fixing one product at a time inside Merchant Center does not scale.
4. Reprocess and verify. After updating rules, regenerate the feed and run end-to-end QA on a sample of affected products. Confirm the fix resolves the issue on the live listing.
5. Track recurrence. If the same root cause reappears the following week, your fix was not durable. Investigate whether the source data reverted or the rule has edge cases.

For a more detailed pre-launch checklist, see our feed QA checklist. And if you are still diagnosing the most common disapproval categories, our disapprovals guide maps each error type to a remediation path.

Teams that follow this cadence consistently find that their disapproval rate drops within weeks and stabilizes at a low baseline. The key insight is that feed optimization is not a one-time fix. It is an operating discipline, and the teams that treat it that way outperform those that react only when Google flags problems.

To accelerate this workflow, explore Lasso features for automated validation and enrichment, review rollout scope on pricing, or schedule a feed audit tailored to your catalog.