Why Your Google Merchant Center Is Missing the Data That Actually Drives Sales

Nobody scrolls ten blue links anymore. A buyer types "best polarized sunglasses for driving under $200" and gets one answer, built by AI, personalized to their history, delivered in seconds. Google AI Mode runs six to ten parallel searches behind that single question. ChatGPT recommends products based on what it knows about the person asking. Perplexity cites specs it can verify. Claude compares on precise attributes. Amazon's shopping AI sells on listing depth.

Every one of these systems makes its decision the same way. It reads structured product data. Not your ad copy. Not your brand story. Not your campaign budget. Your data.

So the question changed. It is no longer whether your product ranks on a results page. It is whether AI systems have enough structured information about your product to recommend it at all. For most enterprise brands, the answer is no. The reason sits inside Google Merchant Center, in two fields almost nobody fills in: product_detail and product_highlight.

How AI Changed What It Means to Show Up in Search

Showing up in search used to mean winning a keyword auction. It now means being readable by machines that build the answer for the buyer before the buyer ever reaches your site.

Two years ago the process was simple. You wrote a decent title, uploaded a clean image, set a competitive bid, and your product appeared. The buyer clicked through, browsed, maybe purchased. That model is going away.

AI Mode does not send buyers to your website. It builds the answer for them. It pulls product specs, compares options, checks pricing, and delivers a recommendation inside the search interface. The buyer never leaves Google. AI Overviews already quote product attributes in their summaries. AI Mode goes further and synthesizes across hundreds of sources to present the single best match for one buyer's constraints.

The personalization is already running. In one academic study of AI assistant memory, 52% of stored entries contained psychological insights about the user. The profile is built. Your data decides whether your product fits it.

This shifts the game from visibility to readability. Your product is not competing for clicks. It is competing for comprehension. The AI has to read your data, understand it attribute by attribute, and match it against what a specific buyer asked for. If your feed is thin, a title, a description, a price, the AI has nothing to work with. It recommends your competitor.

The brands winning here are not spending more on ads. Their product data is rich enough for AI systems to recommend them with confidence. AI does not read your marketing. It reads your data.

The Amazon vs Google Data War, and Why Your Brand Sits in the Middle

Amazon and Google took opposite approaches to product data, and your catalog is caught in the gap between them.

Amazon built its entire experience around mandatory structured data. You cannot sell a laptop without specifying processor, RAM, storage, screen size, and dozens of other attributes. Amazon enforces this because structured data is the engine behind every recommendation, every filter, every "customers also bought" suggestion. The depth is not optional. The depth is the product.

Google did the reverse. Merchant Center requires the minimum: title, description, price, image, availability. Everything else is optional. The theory was that merchants would volunteer richer data. They never did.

Now Google is closing the gap. Free listings launched in 2020 to attract more merchants and more product data. The Google Shopping Graph connects products, reviews, pricing, and inventory across the web. AI Overviews cite specifications. AI Mode synthesizes comparisons. Every one of those features runs on the same fuel: structured product data most merchants are not providing.

Here is the part I find hard to ignore. Google has a data advantage nobody else can match, across Search, YouTube, Images, Maps, and Merchant Center. When a task involves product data, Google's models are training on signals the others simply do not have. Google is handing you the tools to compete with Amazon's depth. The tools are free. They are available right now. Almost nobody picks them up.

That is the opportunity. It lives in two Merchant Center fields.

What Are Product Detail and Product Highlight?

Product highlights are short bullet points that describe the key selling points of a product. Product details are structured attribute-value pairs, organized by section, that describe its specifications. One is for humans scanning fast. The other is for machines comparing precisely.

Think of highlights as the five things a good salesperson says in the first thirty seconds. "Polarized crystal lenses with 100% UV protection." "Lightweight metal frame at just 29 grams." Google shows these in Shopping results and free listings so buyers decide faster. AI systems read them to understand what makes the product different.

Details go deeper. Instead of a sentence, you give specific data points tagged by section.

This is the structured data that powers filters, comparison tables, and every AI-driven recommendation. Together, highlights and details give Google the same depth Amazon gets from its mandatory attribute system. The difference is that Google makes them optional. So most merchants skip them.

I know, because I decided to stop skipping them.

What I Found When I Ran the Experiment on Ray-Ban

The result was clear: enriching structured product data lifted both paid and free performance across the full catalog, with no new ad budget.

I spent two years inside the gap between what Merchant Center can hold and what brands actually provide. Everyone talks about title optimization and bidding strategy. Almost nobody talks about the attribute fields AI systems actually read. So I ran the experiment on a catalog that matters: Ray-Ban.

The approach was direct. I took the full Ray-Ban catalog and enriched every product, not a subset, not a test group, with structured product_detail and product_highlight fields. Lens technology, frame material, UV protection specs, fit dimensions, weight. Every attribute a buyer might filter on or an AI system might use to compare.

The results:

• +12.95% improvement in ad performance across Shopping campaigns
• +5.68% increase in free listing visibility

Those are account-level, full-catalog measurements. Not a cherry-picked product. Not a single campaign. The whole catalog, measured before and after.

What struck me was not the size of the lift. It was how obvious the opportunity had been all along. The data already existed in the product information systems. Lens specs, frame dimensions, material composition, all of it sitting in internal databases. It just was not flowing into Google Merchant Center. Nobody had connected the pipe.

Why Most Enterprise Teams Miss This

Enterprise teams miss these fields because of who owns the feed, not because the data is hard to find. The pattern repeats across catalogs from a few hundred products to tens of thousands.

The Merchant Center usually belongs to the paid advertising team. Their priority is campaign performance, bidding, and ROAS. Feed work beyond titles and images is not on their radar. They treat Merchant Center as an ads platform, not a product data platform.

Free listings get ignored for the same reason. Google made Shopping listings free in 2020, but most enterprise teams still think of Merchant Center as a paid channel only. The free surface, where enriched data matters most, gets overlooked.

The data exists but does not move. Every enterprise has a Product Information Management system, an ERP, and internal spreadsheets full of materials, dimensions, certifications, and care instructions. All collected. All stored. Almost none of it reaches the surfaces where buyers now decide. The asset is already paid for. It just is not working.

Then there is scale. For a catalog of 5,000 or 50,000 products, writing structured details by hand for each SKU is not realistic. Without automation, the task is too big to start.

I will say it plainly. The product_highlight and product_detail fields are the most underused real estate in ecommerce. Platforms like Shopify almost never populate them, which means you own those fields permanently once you write them. Generic feeds give generic results. Systems built around your own data are where the value sits.

How to Do This at Enterprise Scale

Enrich the fields with an automated pipeline that reads your existing product data and writes it into the product_detail and product_highlight format. What used to take weeks of manual entry can run in hours, and it survives every catalog update.

Start with your PIM. The data you need already exists. A "Material" field in your PIM becomes a product_detail with section "Materials," attribute "Primary Material," value "Full-grain Italian leather." The mapping is natural. AI can read that source, whether it sits in a PIM, an ERP, a spreadsheet, or on your own product pages, and structure it into the right format at scale.

One technical trap catches enterprise teams every time: feed overwrites. If your platform, Shopify, Magento, or BigCommerce, already sends a feed to Merchant Center, editing it directly risks breaking what the platform manages. Use a supplemental feed instead. It adds the enrichment without touching your primary feed. Fully reversible. No platform conflicts.

Focus on product_detail, not just product_highlight. Most guides stop at highlights because bullet points are simple. Details are where the value sits.

Highlights are marketing. Details are data. AI reads the data.

What to Do This Week

Open your Merchant Center and check one product. That single action tells you whether you have the gap, and you almost certainly do.

First, look at the product_detail and product_highlight fields on any product in your feed. If they are empty, you found the opening.

Second, audit your data sources. List every attribute your PIM holds that is missing from your Merchant Center feed. Materials, dimensions, specifications, certifications. That list is your enrichment roadmap.

Third, start with one category. Do not try to enrich 50,000 SKUs on day one. Pick your highest-value category, enrich those products with structured details and highlights, and measure the change over 30 days. Use that data to justify the full rollout.

The gap between Amazon and Google is closing. AI systems are choosing which products to recommend based on data depth, not ad spend. The brands that fill their product data first are the ones AI will recommend. Everyone else is hoping titles and images are enough.

They are not. Not anymore.