Product Feed Optimization in Google Merchant Center Next (2026)

In a standard Google Search campaign, you choose your keywords. In Google Shopping — free product listings and paid Shopping ads alike — Google matches your products to queries by reading your product data source (the term Merchant Center Next uses for what was previously called a product feed). There is no keyword field in Shopping. Your product title functions as a broad-match keyword, your description extends coverage into the long tail, your images determine whether clicks follow impressions, and your unique identifiers signal to Google's auction engine that your product is a known, trusted item. Optimizing the product data source is therefore the core optimization activity in Shopping — not bid management, not audience layering, not ad copy. The data is the campaign structure.

This guide covers the five levers that move Shopping performance in Merchant Center Next: product titles (the 150-character spec, the ~70-character visible window, and the vertical-specific attribute order), product descriptions (the 5,000-character limit and why the first 160–500 characters carry disproportionate weight), product images (the full spec table — current minimums, the incoming January 31, 2027 universal 500×500 px enforcement, and the recommended 1,500×1,500 px threshold Google uses internally), feed quality (and the myth-correction: Google publishes no feed quality score — the actual signals are product status, the Needs attention view, and the separate Top Quality Store program), and Product Studio (Google's free AI image and text tool in 2026, with its exact region limits). It also covers unique identifiers (GTINs), listing visibility, custom labels, seasonal timing, and the Preemptive Item Disapproval mechanic.

All merchants migrated to Merchant Center Next by September 2024. Classic Merchant Center flows — the Diagnostics tab, the Feeds section, the Feed Rules page — no longer exist in the same form. The current terms are data sources (not feeds), attribute rules (not feed rules), and Needs attention (not Diagnostics). Throughout this guide, current Merchant Center Next terminology is used; legacy terms are noted where confusion is common.

Before optimizing, the foundation is understanding what a data source actually contains — required attributes, unique identifiers, and how supplemental sources work. See what is a product feed in Google Merchant Center for that baseline. Optimization is only as good as your refresh cadence — for scheduling, attribute rules, and automatic item updates, see managing and automating product feeds. For fixing active disapprovals — price and availability mismatches, invalid GTINs, and image violations — see debugging feed disapprovals and errors in Merchant Center. The programs layer that turns feed quality into incremental reach is covered in overview of Merchant Center programs.

At MB Adv Agency, our Shopping feed management team works with merchants across ecommerce verticals from fashion and furniture to beauty and electronics. The pattern we see consistently: merchants who fix their titles to front-load brand and key attributes, correct their images to the recommended 1,500×1,500 px resolution, and add correct GTINs to their top-selling products see the highest-impact gains — before any bid adjustment. The feed is the lever. This guide covers how to pull it.

Key Takeaways

• Product data is the keyword system in Shopping. There are no keyword fields. Google reads your title, description, and attributes to match products to queries. Optimizing the data source is the primary campaign optimization activity.
• Titles: 150-character max, ~70 characters shown. Front-load brand + product type + key differentiator in the first 70 characters. Use vertical-specific attribute order (apparel: Brand → Gender → Product Type → Color → Size). No ALL CAPS, no promotional text — these trigger disapproval or Google auto-rewrite via Product Data Customization.
• Descriptions: 5,000-character max, plain text only. Put the most important information in the first 160–500 characters. Google recommends 500–1,000 characters for practical optimization. No HTML, no links, no promotional text.
• Images: recommended 1,500×1,500 px or larger. The current minimums are 100×100 px (non-apparel) and 250×250 px (apparel). A new universal 500×500 px minimum is enforced from January 31, 2027, with warnings active since April 14, 2026. The "800px recommended" figure is not in current Google documentation. No promotional overlays on the primary image_link.
• Feed quality score: it does not exist. Google publishes no single feed quality score. The real signals are product status (Approved / Limited / Not Approved / Under Review / Processing) and the Needs attention view. Merchants chasing a score that is not in Merchant Center have been looking at the wrong metric.
• GTINs: +20% clicks on average (Google stat). Google reports that products with correct GTINs see a 20% increase in clicks on average. GTINs are free to add from manufacturer spec sheets. Use identifier_exists: no only for genuinely identifier-less goods.
• Product Studio is free and built into Merchant Center Next. AI image generation, background swapping, resolution upscaling, video animation (AU/CA/IN/JP/UK/US), and AI title and description suggestions. Not available in Russia or France.

Product Descriptions: Specifications and Best Practices

The description attribute is the long-tail extension of your title. Google reads it to understand product context beyond what the title covers, and it surfaces in some ad formats and the free listing detail page. The five-thousand-character maximum sounds generous, but the field has hard formatting and content constraints that catch merchants off-guard.

Write the description as if the first sentence is the only sentence a shopper will read. Lead with the product's function, its key specification or differentiating ingredient, and the category — not the brand story. The brand is already in your title and the brand attribute. Descriptions are also where keyword depth for the long tail lives: a title for a camera lens covers the model and aperture; the description covers "full-frame compatible," "weather-sealed," and "macro capable" — the qualifier queries that drive informed buyers.

Product Images: Google Merchant Center Recommended Image Size and Requirements (2026)

Google recommends 1,500×1,500 pixels or larger for all products. This is the resolution Google uses to serve high-quality placements across all Shopping formats. Images below 1,024 px fall below Google's high-resolution threshold. The widely-cited "800px recommended" figure is not in current Google documentation — it is a drift trap from older guides.

Two spec changes are active right now and one is incoming. The current enforced minimums are 100×100 px for non-apparel and 250×250 px for apparel. A new universal minimum of 500×500 px takes effect January 31, 2027 — Google began issuing warnings in Merchant Center's Needs attention view on April 14, 2026. Merchants currently meeting the 100/250 floor but submitting images below 500×500 are already seeing warnings and will receive disapprovals after the enforcement date. Google also announced it would auto-upscale some images using AI as a safety net — but that is not a substitute for submitting compliant images.

Image violations are consistently the most common and most avoidable disapproval category in the merchant feeds reviewed by MB Adv Agency's Shopping team. The two primary failure modes: images below the resolution floor (now 100/250, soon 500) and promotional overlays on the primary image_link image — both cause immediate disapproval with no grace period.

Product Studio's image-upscaling tool is a practical solution for product images that are compliant now but will fall below the January 2027 500×500 floor — it is free and built into Merchant Center Next. For lifestyle or seasonal creative, use additional_image_link (where scene images and overlays are permitted) and keep the primary image_link clean.

Product Titles: Vertical-Specific Structure and Prohibited Content

Every thin optimization guide says "put keywords in your title." The instruction that actually works in Shopping is: front-load the attributes that matter most for your vertical, in the order shoppers use to filter. The 150-character title limit is generous, but only ~70 characters are shown in standard Shopping results. The first 70 characters must contain brand, product type, and the most differentiating attribute — not "Shop our amazing collection of..." The table below documents the vertical-specific attribute order Google's product data specification recommends, with the most common disapproval-triggering errors by vertical.

In a January 2025 DataFeedWatch case study, United RV applied AI-generated title rewriting to 3,500 Google Shopping products, rebuilding every title to front-load brand, product type, and key attributes per Google's best practices. Category A saw a +17% CVR increase and Category B saw a +44% CVR increase over two weeks. Improvements were visible "the very next day after pushing AI-generated titles live" (DataFeedWatch). These are DataFeedWatch-published figures for one RV merchant; see the DataFeedWatch case study for full methodology.

Google Product Data Customization rewrites titles automatically when it detects non-compliance or thin content. If your titles are non-compliant, Google's rewrite is not guaranteed to represent your product accurately. The practical implication: own your titles before Google does. Use a supplemental data source with attribute rules to apply title transformations at scale without touching your primary data source — the workflow is covered in managing and automating product feeds.

The "Feed Quality Score" Myth: What Feed Quality Actually Looks Like in Merchant Center Next

Google does not publish a feed quality score. This is the most-searched metric in Merchant Center optimization — and it does not exist as a Google-published number. The term "feed quality score" originates from third-party feed management vendors and the SEO content ecosystem. It has never appeared in Merchant Center's own interface. Merchants who have been chasing a score they cannot find are looking for something that is not there.

What Merchant Center Next actually shows is a set of distinct signals: per-item product status, the Needs attention view (item-level issues ranked by click potential), and the separate Top Quality Store program — which scores the merchant's purchase experience, not the feed data. The table below maps each signal to its location in the Merchant Center Next UI and what to optimize against it.

The practical quality target is simple: zero disapprovals, zero High click-potential issues in Needs attention, and Approved status on every revenue-driving product. That is the Merchant Center Next equivalent of a "perfect feed quality score" — expressed in the signals Google actually publishes.

Google Product Studio: AI Capabilities, Availability and Limits (2026)

Product Studio is free, built into Merchant Center Next, and covers image generation, resolution upscaling, video animation, and AI-generated titles and descriptions. It is not a replacement for professional product photography — especially for high-end catalogs where brand-consistent imagery drives purchase decisions. It is a practical tool for generating lifestyle scene variants from a single product shot, upscaling images that are approaching the upcoming 500×500 minimum, and templating seasonal creative without a full creative production cycle.

Two availability limitations are confirmed as of 2026-06-30 (exact Google quote from answer/13708167): "Product Studio is available to all Merchant Center users except those based in RU and FR." And for video generation: "The Generate Video feature is only available for merchants in Australia, Canada, India, Japan, the UK, and the US."

Product Data Optimization Checklist (Google Merchant Center Next, 2026)

The checklist below organizes the optimization work covered in this guide by priority: Critical items cause immediate disapprovals with no grace period; High items affect the broadest set of queries and impressions; Medium items are important but not immediately blocking; Low items are campaign-level benefits with no Shopping ranking impact. Run through this list in priority order before turning on or scaling any Shopping or Performance Max campaign.

Listing Visibility: Unique Identifiers, Product Category, and Variant Attributes

Feed quality is necessary but not sufficient for Shopping visibility. Three attribute groups determine whether your products surface in the right queries at the right frequency: unique identifiers (GTINs signal a known product to Google's matching engine), google_product_category (precision here directly affects query-to-product match accuracy), and variant attributes (which prevent duplicate product cards and enable size/color filtering in results). Custom labels unlock campaign segmentation — they have no effect on organic visibility but are essential for Shopping and Performance Max bid strategy.

Systematic identifier and category work across a large catalog is the kind of optimization that a Google Shopping feed management engagement prioritizes first — it lifts the floor for the entire catalog rather than optimizing individual products. For a full audit of identifier coverage and category precision in your existing data source, an ecommerce PPC audit surfaces the gaps before you run them.

Seasonal Feed Readiness: What to Have Live Before Your Peak Window

Price and availability accuracy failures are the most common cause of Shopping disruption during high-traffic windows. A Preemptive Item Disapproval during peak traffic is a lost opportunity with no recovery time — review periods during Black Friday or holiday weekends are not accelerated. The seasonal optimization discipline is: get your feed, promotions data source, and price history synchronized weeks before the peak, not days. The table below provides a T-minus framework — anchored to your peak window, not to calendar dates, since different verticals have different peaks.

What Product Feed Optimization Actually Delivers (Sourced Evidence)

Google's own product data specification puts one number on the table for identifier optimization: "Retailers who've added correct GTINs to their product data have seen a 20% increase in clicks on average" (Google Merchant Center, Tips to optimize your product data, confirmed 2026-06-30). This figure is an average across all product categories. Individual category results differ. The stat is attributable to Google — it is not mbadv client data and not a benchmark fabricated from partial information.

For title optimization, the most specific named-source case data available is a January 2025 DataFeedWatch case study on AI title rewriting. United RV applied DataFeedWatch's AI title optimization to 3,500 Google Shopping products, rebuilding every title to front-load brand, product type, and key attributes per Google's best practices. Results after two weeks: Category A saw a +17% CVR increase; Category B saw a +44% CVR increase. Improvements were visible "the very next day after pushing AI-generated titles live" (DataFeedWatch). These are DataFeedWatch-published figures for one RV/camper merchant across one two-week period — not a universal benchmark and not mbadv client data. See the full DataFeedWatch AI Feed Optimization Case Study for methodology. Only CVR lift was disclosed; no CPC or impression-volume data were reported.

The two external data points above are from different optimization levers (identifiers vs. titles), measured differently (click volume vs. conversion rate), and cannot be combined into an aggregate "feed optimization impact" number. What they share is the mechanism: better product data → more accurate query-to-product matching → more relevant impressions → more purchase-intent clicks. The feed is the signal Google reads to decide whether your product is the right answer for a given search. Better data means more correct answers.

For ecommerce merchants across verticals from beauty and supplements to furniture and electronics, the MB Adv Agency Shopping feed team treats product data optimization as the prerequisite to bid strategy — not the follow-up. A campaign running on incomplete or non-compliant data is not a campaign that benefits from bid changes. The feed work comes first.

Five Misconceptions That Cost Merchants Shopping Performance

1. "I have a good feed quality score." This is the most common myth in Merchant Center optimization. Google publishes no feed quality score — the term does not exist in Merchant Center Next. Third-party feed management vendors use it to describe their own internal scoring systems, not a Google metric. The actual proxies to track are: zero Not Approved products on revenue SKUs, zero High click-potential issues in Needs attention, and Approved status across the catalog. That is what Google shows; there is no number behind it.

2. "The recommended image size is 800px or 1,024px." Both figures circulate widely on older guides but neither is in current Google documentation. Google's current recommendation is 1,500×1,500 px or larger. The 1,024 px figure represents Google's high-resolution threshold — images above it are classified as high-resolution, not as the recommended size. The 800px figure has no current Google source. Use 1,500×1,500 px or larger now; it future-proofs against the January 31, 2027 universal 500×500 px enforcement.

3. "Putting 'Free Shipping' and 'SALE' in the title attracts more clicks." Google's product data specification explicitly prohibits promotional language, ALL CAPS text, and gimmicky characters in the title attribute. These trigger disapproval or automated title replacement via Product Data Customization — which means Google rewrites your title with its own version, which is not guaranteed to represent your product accurately. Promotional offers belong in the Promotions data source (separate add-on), where they display as sale badges beneath Shopping listings without touching the product title.

4. "I don't need GTINs for my private-label products." This is correct for genuinely identifier-less goods (custom, handmade, vintage, private-label products with no manufacturer-assigned GTIN). For those, identifier_exists: no is the right submission. But merchants who skip GTINs on branded products that do have manufacturer-assigned GTINs are forfeiting what Google describes as a 20% click increase on average — and generating a warning that leads to disapproval. The check is simple: does the manufacturer publish a GTIN for this product? If yes, submit it.

5. "Feed rules" and "feeds" are the current Merchant Center terms. All retailers migrated to Merchant Center Next by September 2024. The current terms are data sources (not feeds) and attribute rules (not feed rules). Guides that instruct you to "go to your feed settings" or "set up feed rules" describe legacy Classic Merchant Center navigation that no longer exists in the same form. Supplemental data sources and attribute rules are accessed differently in Merchant Center Next — and are the correct workflow for attribute-level overrides without touching the primary data source.

Frequently Asked Questions

Methodology

This pillar consolidates seven absorbed Google Merchant Center pages covering product titles, descriptions, images, listing visibility, feed quality signals, Product Studio, and seasonal optimization. Every specification, status label, and product attribute in this guide traces to Google's official Merchant Center documentation (support.google.com/merchants) or the shared verified-facts baseline, web-verified June 30, 2026. The GTIN "+20% clicks on average" is Google's own cited figure from answer/7380908 — attributed to Google at every usage. The DataFeedWatch CVR lift figures (+17%, +44%) are from the DataFeedWatch AI Feed Optimization Case Study (January 2025) for United RV — a single merchant case study, not a cross-industry benchmark.

No mbadv client metrics. MB Adv Agency has no proprietary Google Shopping feed performance dataset. Agency POV in this guide is qualitative: "the most common disapproval category we see," "when we audit merchant feeds." Every quantified claim is attributed to Google or a named third party. No WordStream or LocaliQ Shopping cost figures appear anywhere in this guide — their only Shopping figure is a stale 2019 CPC estimate not applicable to 2026 optimization benchmarks.

Verify before publish: The April 14, 2026 warning-start date and January 31, 2027 enforcement date for the universal 500×500 px image minimum are load-bearing — re-confirm against the live Google Merchant Center image requirements page (answer/6324350) and product data specification update (answer/16989427) at publish. The "feed quality score does not exist" statement is load-bearing — confirm no Google-published feed-level quality score has been introduced. Product Studio capabilities evolve frequently — re-verify at publish. Last updated: 2026-06-30.