Ecommerce SEO / AI Shopping / Google Merchant Center
Google Merchant Center AI Performance Insights: 2026 Guide
Learn how to access Google's AI performance report, understand your brand's share of voice, investigate product visibility gaps, and turn Merchant Center data into accurate product-feed improvements.
What are Google Merchant Center AI Performance Insights?
Google Merchant Center AI Performance Insights is a reporting feature that helps merchants understand how their brands and products appear in conversational shopping queries on Google AI Mode and AI Overviews. The report includes share of voice, competitor comparisons, shopping-stage performance, popular search terms, and product attributes. Merchants can use these insights to identify product-data gaps and investigate opportunities to improve visibility.
Low AI share of voice does not automatically mean your product feed is incorrect. The report provides a signal for investigation, but you need to check your actual catalog, product specifications, Merchant Center data, and landing pages before deciding which changes are justified.
Understand AI visibility
Review share of voice, shopping stages, and relevant product terms.
Identify catalog gaps
Compare the reported shopping need with your actual product specifications.
Correct product data
Update missing or inconsistent information using verified product records.
Measure subsequent results
Confirm feed processing and compare performance using consistent reporting settings.
1. How to Access Google Merchant Center AI Performance Insights
Google introduced AI Performance Insights to help merchants understand how their products and brands appear during conversational shopping journeys.
As of September 2026, the report is available for English-language queries to eligible Merchant Center accounts in Australia, Canada, India, New Zealand, and the United States.
To access your report, sign in to Google Merchant Center and follow this navigation path:
Google's official Merchant Center announcement includes an illustration of the AI Performance Insights dashboard with visibility measurements and a shopping-stage breakdown.
The illustration uses simulated results and fictional entities. It is useful for understanding the interface, but its figures are not actual merchant performance data.
View Google's Original Dashboard IllustrationWhich filters should you select?
Before analyzing performance, choose the product category, country, and reporting period relevant to your investigation.
Merchant Center does not provide a single AI performance report covering all product categories. Analyze the appropriate category rather than assuming that one category represents your entire catalog.
The report focuses on organic AI traffic, including free listings. Paid advertising traffic is not included.
For the current access instructions and reporting requirements, consult Google's official AI Performance Insights documentation .
2. How to Interpret AI Share of Voice and Other Performance Metrics
Merchant Center's AI performance report contains four important measurements that help merchants understand relative product and brand visibility.
These measurements answer different questions. They should not be treated as interchangeable.
| Metric | What it measures | What to investigate |
|---|---|---|
| Your share of voice | Your brand's share of AI impressions within the defined competitor comparison. | Relative visibility for the selected shopping context. |
| Competitors' average share | Average share of voice across the defined competitor set. | How your reported visibility compares with the available benchmark. |
| Frequency | The relative popularity of a reported search term, shopping intent, or attribute. | Which relevant customer needs deserve further investigation. |
| Products showing | The number of your products appearing within a reported shopping context. | Whether your catalog and submitted data adequately describe relevant products. |
What does your AI share of voice mean?
Your share of voice represents the proportion of AI impressions your brand receives relative to the combined impressions for your brand and the defined Merchant Center competitor set.
For example, consider a fictional reporting scenario in which your brand receives 150 AI impressions and the combined total for your brand and the competitor set is 1,000.
Understanding a 15% AI share of voice
Your share of voice would be:
This is an original mathematical illustration, not real merchant-account performance data.
The 15% figure describes relative visibility within the defined comparison. It does not mean that the brand appeared in 15% of every shopping conversation on Google.
It also does not establish a 15% market share, click-through rate, or purchase rate.
SEO consultant Brodie Clark shared a screenshot from a merchant account showing actual historical AI performance figures. The screenshot was subsequently discussed and reproduced by SEO Südwest.
The original publication allows readers to examine the dashboard and its reported metrics. The data belongs to a third-party merchant account and does not represent SearchCounselCo's own search performance.
View the Real Merchant Center ScreenshotHow should you investigate a low share of voice?
Start by checking whether the reported shopping term or attribute matches products your store actually sells.
Then review the relevant product category, shopping stage, available competitor benchmark, and products showing count.
A low share of voice may justify a catalog investigation, but the metric alone does not identify an incorrect product title or a specific missing feed field.
Is frequency the same as Google keyword volume?
No. Frequency represents the popularity of reported terms or attributes within the AI performance report.
It is not a direct measurement of monthly Google keyword search volume.
Use it to prioritize shopping needs within the report rather than presenting it as conventional SEO keyword-volume data.
3. Understanding Discovery, Evaluation, and Ready-to-Buy Performance
Google classifies conversational shopping queries into three shopping stages.
These stages can help merchants understand which product information shoppers need at different points in their buying journey.
Discovery
Shoppers explore product categories, possible solutions, and general features.
Example: "What type of running shoes should someone with wide feet look for?"
Investigate: Product categories, intended uses, features, and accurate descriptions.
Evaluation
Shoppers compare products and consider specifications, alternatives, and suitability.
Example: "How do wide-fit running shoes compare with regular-fit models?"
Investigate: Verified specifications, sizing details, materials, and product differences.
Ready to buy
Shoppers investigate specific products and purchasing options.
Example: "Where can I buy Model A in a men's wide-fit size 10?"
Investigate: Variant accuracy, availability, price, and landing-page consistency.
Match the real product
Do not add every popular attribute to every product description.
Only describe features that the product genuinely has.
For example, a regular-fit shoe should not be described as wide fit merely because wide-fit queries are popular in the AI performance report.
Your product information should answer real customer questions while remaining consistent with verified specifications.
4. How to Audit Low Product Visibility in Google AI Shopping
The most useful application of AI Performance Insights is connecting aggregate reporting signals with the actual information available in your product catalog.
The report can identify relevant shopping terms and product attributes. It does not, by itself, establish the exact underlying reason for every visibility difference.
Use the following six-step process to investigate possible product-data issues.
Record the reporting context
Select one relevant product category, country, and time period.
Record the shopping stage, reported term, frequency, share of voice, competitor benchmark, and products showing count.
Choose a term that corresponds to a genuine customer need and is relevant to products your business sells.
Identify candidate products
Search your own catalog for products that genuinely match the reported shopping need.
Record the product IDs, titles, specifications, variants, and landing-page URLs.
Verify the relevant features against manufacturer information or other reliable internal product records.
This creates a list of products to investigate. It does not establish that those exact products appeared in Google's aggregate report.
Compare catalog records with the feed
Review the information currently submitted to Merchant Center.
Check the product title, description, product details, variant attributes, identifiers, price, availability, and landing-page URL.
Look for information that is missing, incorrect, outdated, or inconsistent.
Do not change an accurate product record simply because it does not repeat a popular search term.
Review the corresponding product page
Open each candidate product's landing page and confirm that its visible information matches the actual product and the submitted product data.
Check the model, features, specifications, variant options, price, and availability.
If a verified product feature is missing from the page, add useful information in an appropriate location.
For additional guidance, read our product page SEO guide .
Make the verified correction
Update the appropriate product field when the audit identifies a genuine information gap.
A verified wide-fit shoe variant, for example, may require clearer width information in its product title and variant attributes.
Preserve correct product identifiers and existing variant relationships.
If the product information is already complete and accurate, document that finding rather than making unnecessary changes.
Validate and measure the outcome
Confirm that Merchant Center has received and processed the corrected product data.
Review any relevant processing or approval issues and confirm that the landing page displays matching information.
Record the implementation date, the fields changed, and the original reporting measurements.
Compare subsequent performance only after the relevant data has refreshed.
5. Worked Example: Correcting a Missing Product Attribute
The following retailer, products, identifiers, and feed records are fictional. They demonstrate a reproducible audit methodology and are not actual client performance results.
The initial AI visibility opportunity
Imagine an online footwear retailer whose Merchant Center report shows frequent interest in wide-fit running shoes.
The retailer has relatively low reported visibility for that shopping context.
Rather than adding "wide fit" to every running-shoe listing, the retailer checks which products actually have a verified wide-fit variant.
Step A. Investigate the catalog
| Product ID | Verified specification | Audit finding |
|---|---|---|
| SHOE-A-W10 | Model A, men's size 10, wide fit. | The title does not clearly identify the width variant. |
| SHOE-A-R10 | Model A, men's size 10, regular fit. | The product does not support a wide-fit claim. |
| SHOE-B-W09 | Model B, men's size 9, wide fit. | The width is already represented correctly in the product data. |
Only the first product has a confirmed information gap in this example.
That conclusion comes from the retailer's catalog audit, not from an individual product ID provided by the aggregate AI report.
Step B. Correct the verified information
Incomplete variant information
Product ID: SHOE-A-W10
Title: Model A Running Shoes
Size: 10
Width: Not clearly represented.
Description: Lightweight running shoes with comfortable cushioning.
Accurate variant information
Product ID: SHOE-A-W10
Title: Model A Wide-Fit Running Shoes, Men's Size 10
Size: 10
Width: Wide
Description: Model A running shoes in the verified men's wide-fit version, size 10, with the model's documented cushioning and fit specifications.
Step C. Demonstrate the product-feed correction
The retailer verifies the product specifications and updates the appropriate source fields.
The following example illustrates the relationship between standard product data and optional conversational variant attributes.
Illustrative product-data example only. It is not a complete production feed. The example URL and product identifiers are fictional. Submit all other required attributes using the correct format for your data-source type.
Google's optional
variant_option
attribute can describe properties
that distinguish variants within
a product group.
It should be used with the appropriate item group information, and the variant properties must remain consistent across related records and their landing pages.
Consult Google's variant option specification for formatting and implementation requirements.
Step D. Validate the corrected record
After submitting the updated data, the retailer confirms that the relevant product record has been processed in Merchant Center.
The landing page should describe the correct wide-fit variant, and any applicable product structured data should identify the same product.
Merchant Center product feeds and on-page structured data are different data sources, but both should provide consistent information.
For more information, see our product structured data guide .
Step E. Measure subsequent visibility
The retailer records the original category, country, reporting period, shopping term, shopping stage, share of voice, and products showing count.
After the corrected feed has been processed and the report has refreshed, the retailer compares the same reporting context.
A subsequent increase in visibility is an observed change. It does not automatically establish that the single feed correction caused the increase.
The merchant uses an AI visibility signal to investigate its actual catalog, identifies a genuine information gap, corrects it, validates the implementation, and measures subsequent results.
Even if visibility does not increase, the retailer has improved the accuracy of its product information.
6. Which Product Attributes Should You Optimize for AI Shopping?
Google recommends maintaining high-quality product data, incorporating relevant shopping terms into appropriate titles and descriptions, and completing missing attributes.
Start with accurate core product information rather than adding optional fields without a clear reason.
Review existing product data first
Check the product title, description, product details, identifiers, price, availability, and variant information.
Your product feed and landing page should accurately describe the same product being sold.
When should you use conversational attributes?
Google supports six optional conversational attributes that can provide additional product information for AI-supported shopping experiences.
| Attribute | Potential use |
|---|---|
question_and_answer
|
Provide verified answers to genuine product-specific questions. |
document_link
|
Provide relevant product documents, such as an authoritative PDF manual. |
related_product
|
Describe supported relationships between products and accessories. |
item_group_title
|
Identify the shared name of a product group containing multiple variants. |
variant_option
|
Describe properties that distinguish individual product variants. |
popularity_rank
|
Provide supported product-popularity information using Google's specified format. |
These attributes are optional. They complement the primary Merchant Center product data specification.
Google recommends using a supplemental data source, although supported attributes can also be submitted through a primary data source or the Merchant API.
Follow the individual attribute specifications and do not duplicate information unnecessarily when it is already available in the existing product data.
For the complete requirements, see Google's conversational attributes documentation .
No. Adding optional product attributes does not guarantee inclusion in AI shopping results, more impressions, or additional sales. The information must be accurate and genuinely useful for understanding the product.
7. AI Performance Insights Not Showing? Troubleshooting Guide
If the report is missing or displays unexpected values, investigate eligibility, reporting filters, and data availability before changing product information.
The AI performance tab is missing
Confirm that the account is in an eligible country and follow Analytics → Products → AI performance. If the report remains unavailable, consult the current Merchant Center documentation and support resources.
Your AI share of voice displays 0
Google states that a share-of-voice value of 0 can indicate insufficient AI impressions. Check the reporting context and available data before concluding that a feed correction is required.
Your share of voice displays a dash
A dash indicates that no impressions data is available for the reported context. It does not independently establish a ranking penalty or product-data error.
Your share of voice displays 100%
Google states that 100% can appear when the account has insufficient defined competitor data. It does not necessarily mean the brand appears in every relevant AI shopping result.
Products showing displays 0
A value of 0 indicates that no products are showing within that reported context. Check whether your catalog contains relevant products and whether the corresponding information is accurate.
Your visibility changes unexpectedly
Compare equivalent reporting periods, countries, categories, and shopping contexts. Google notes that changes in the defined competitor data can affect comparative share-of-voice measurements.
Historical reporting data updates daily with a lag of several days. Avoid drawing conclusions from an incomplete recent period.
8. Merchant Center AI Performance vs Google Search Console
Merchant Center and Search Console measure different aspects of visibility in Google's search and shopping experiences.
Merchant Center AI Performance Insights focuses on how brands and products appear within supported conversational shopping queries.
Search Console focuses on the visibility and performance of website URLs.
| Question | Merchant Center AI Performance | Search Console |
|---|---|---|
| Primary focus | Brand and product visibility in conversational shopping. | Website URL visibility and search performance. |
| Competitor comparison | Reports share of voice against a defined competitor set. | Does not provide an equivalent Merchant Center brand share-of-voice measurement. |
| Product information | Reports shopping terms, attributes, and products showing. | Focuses on website URL performance rather than product-feed attribute completeness. |
| Individual website URLs | The AI report emphasizes aggregate product and brand visibility. | Provides page-level search performance data. |
| Purchases and revenue | Share of voice alone does not establish sales. | Search visibility alone does not establish sales. |
Google also introduced dedicated generative AI performance reports in Search Console in 2026. These provide additional visibility information for supported AI search features.
Consult Google's Search Console generative AI reporting announcement for the available measurements and reporting dimensions.
Use Merchant Center to investigate product and brand visibility in supported AI shopping queries.
Use Search Console to analyze website URL visibility and search performance.
Use your ecommerce analytics to investigate recorded purchases, revenue, and other business outcomes.
Do not combine impressions from different systems unless their measurement scopes and potential overlap are understood.
9. How to Measure Product Visibility After Making Changes
Product-data corrections and subsequent visibility changes are separate observations.
A reliable measurement process should record the original information gap, the correction, the implementation date, and the results observed afterward.
Product visibility audit worksheet
Use the following worksheet structure in your spreadsheet or ecommerce reporting system.
| Audit field | Information to record |
|---|---|
| Product category | Category selected in Merchant Center. |
| Country and reporting period | Filters used for the baseline. |
| Shopping stage | Discovery, evaluation, or ready to buy. |
| Reported term or attribute | The shopping need being investigated. |
| Baseline frequency | Available frequency measurement. |
| Baseline share of voice | Reported relative brand visibility. |
| Competitor benchmark | Available competitor comparison. |
| Products showing | Aggregate count in the report. |
| Candidate product ID | Product identified through your catalog audit. |
| Verified specification | Evidence supporting the product claim. |
| Information gap | Specific missing or inconsistent information. |
| Correction | Exact field and value changed. |
| Implementation date | Date the source information was updated. |
| Validation status | Processing and consistency checks. |
| Follow-up measurements | Subsequent comparable performance data. |
| Interpretation | Observed changes and remaining uncertainty. |
Which performance metrics should you track?
AI visibility
Review share of voice, popular terms, and products showing.
Listing engagement
Review available product impressions, clicks, and relevant listing performance.
Website visibility
Investigate relevant URL performance and available AI search reporting metrics.
Business outcomes
Measure recorded product engagement, purchases, and revenue.
When should you review the results?
First confirm that the corrected data has been processed in Merchant Center.
Then allow the relevant AI performance reporting period to refresh.
Compare equivalent categories, countries, shopping contexts, and time periods.
For categories with limited data, use longer comparable periods rather than drawing conclusions from small daily fluctuations.
What if visibility does not improve?
Confirm that the correction was successfully implemented and that the underlying shopping need remains relevant to the catalog.
Investigate changes in availability, competing products, catalog coverage, and reporting conditions.
Do not assume that another product-title rewrite is necessary simply because share of voice has not increased.
A corrected feed improves the accuracy of the submitted product information.
An observed visibility change is a separate outcome. Establishing whether the correction caused that change requires additional evidence.
10. Frequently Asked Questions
What is Google Merchant Center AI Performance Insights?
It is a reporting feature that helps eligible merchants understand brand and product visibility in conversational shopping queries on Google AI Mode and AI Overviews.
Where can I find the AI performance report?
Sign in to Merchant Center, open Analytics, select Products, and choose the AI performance tab.
Which countries support AI Performance Insights?
As of September 2026, Google documents availability for English-language queries to eligible accounts in Australia, Canada, India, New Zealand, and the United States.
Why is AI Performance Insights not showing?
Check account eligibility and the current Merchant Center navigation path. If the report remains unavailable, consult Google's documentation and support resources.
Does AI share of voice measure sales?
No. Share of voice measures relative visibility within a defined comparison. It is not a direct measurement of purchases, revenue, or overall market share.
Can I identify every individual product appearing in AI results?
The AI performance report provides aggregate measurements, including products showing counts. Use your own catalog and verified product information to investigate candidate products rather than assuming the aggregate count identifies every individual product.
Do conversational attributes guarantee AI visibility?
No. Optional conversational attributes can supply additional product information, but they do not guarantee AI visibility, clicks, or sales.
How frequently does the report update?
Google states that historical data is updated daily with a lag of several days. Account for this delay when comparing recent performance periods.
The One Thing to Do Next
Open Merchant Center AI Performance Insights and identify one frequently reported shopping term or product attribute relevant to your catalog.
Before changing your product feed, verify which products genuinely match that shopping need.
Compare their specifications with the information submitted to Merchant Center and displayed on their landing pages.
If you identify a genuine information gap, correct it, validate the implementation, and establish a baseline for measuring subsequent performance.
This turns AI performance reporting into a documented ecommerce SEO investigation rather than a series of speculative keyword changes.
Sources and Further Reading
-
Google Merchant Center:
About AI Performance Insights
Official report navigation, availability, metrics, shopping stages, optimization recommendations, and reporting limitations.
-
Google Merchant Center:
Insights for AI-Powered Shopping Experiences
Original feature announcement and official dashboard illustration. The published example uses simulated results.
-
Google Merchant Center:
Conversational Attributes
Optional product attributes, submission methods, and implementation guidance.
-
Google Merchant Center:
Variant Option Specification
Formatting and implementation requirements for optional product variant information.
-
Google Search Central:
Generative AI Performance Reports
Dedicated generative AI visibility reporting in Search Console.
-
SEO Südwest:
Real Merchant Center AI Reporting Example
Third-party reporting and a historical merchant-account screenshot originally shared by Brodie Clark.
Editorial note: This guide uses Google's publicly available documentation reviewed in September 2026. The footwear retailer, product identifiers, feed examples, and mathematical illustrations are fictional. They are not SearchCounselCo client performance results.
Screenshot attribution: The official Google dashboard illustration is linked to its original publication. The historical merchant-account screenshot is linked to the third-party publisher. Neither third-party image is reproduced without confirmed permission.