Product Page SEO Design

Tomohiro Iida (CEO, Netsujo Inc.) · Published June 26, 2026 · Updated June 26, 2026

Most underperforming e-commerce product pages have the same root cause: information design, not technical SEO. This guide, written for EC and D2C operators, works through the four typical causes — reused manufacturer copy, thin descriptions, mass-produced duplicate variant pages, and structured-data misuse — then covers how to write to actual search intent, how to implement Product/Offer structured data correctly (marking up only what the page actually displays, since anything else is a violation of Google's guidelines), how to decide between merging or splitting color/size variants and when canonical actually applies, how images and original, first-hand product information build differentiation, what makes product text legible to AI systems, and a measurement and improvement cycle using Search Console. Throughout, the article is explicit that none of this guarantees rankings, rich results, AI citation, or Shopping placement — it improves the odds of being evaluated correctly, nothing more.

Why product pages go unnoticed by search and AI

• Reused manufacturer copy — pasting a manufacturer's or wholesaler's stock description verbatim, which then appears near-identically across many other sites, giving search engines and AI little to distinguish and little reason to cite it as a source.
• Thin descriptions — pages with little beyond a name, model number, and price leave shoppers (and AI) without answers to "does this fit my use case," "how is it different," or "what does the sizing feel like."
• Mass-produced duplicate pages — separate URLs for each color or size with near-identical copy make it hard for a search engine to pick a representative page, splitting the evaluation; abandoned URLs for discontinued or out-of-stock items compound the problem.
• Structured-data misuse — marking up a price or stock level that does not match what the page displays, or a review count the page never shows.

Writing to search intent

• Branded search — looking for a specific named product ("[brand] [product name]").
• Model-number search — checking exact specs or stock using a model number.
• Comparison search — weighing multiple candidates ("[category] recommended," "[product] difference").
• Use-case search — searching from a goal or scenario ("[use case] [category]").

Layer original, first-hand information on top of manufacturer copy: the use case or scene, sizing and hands-on feel with actual measurements, notes and caveats beyond the spec sheet, how the product differs from similar or competing items, and answers to common pre-purchase questions. This first-hand information is what gives search engines and AI something to evaluate or cite.

Getting Product/Offer structured data right

The one absolute rule: only mark up what the page actually displays. Writing a price, stock level, or review count into structured data that the page does not show, or that does not match reality, is a violation of Google's structured-data guidelines and can result in a manual action. The Offer type's three core properties are price (must match the displayed price), priceCurrency (JPY for yen), and availability (e.g. https://schema.org/InStock or https://schema.org/OutOfStock, matching what the page shows). aggregateRating or review markup should only be added if the page actually displays reviews, and must match the displayed count and rating. Validate the markup with Google's Rich Results Test or the Schema.org validator before publishing — passing validation still does not guarantee display.

Handling duplicates and variants

• Merge variants into one page when descriptions and specs are shared and switching color/size within the page feels natural — this concentrates evaluation on a single URL.
• Split variants into separate pages when there is clear search demand for a specific variant (e.g. "[product] [color]") — but watch for duplicate content as descriptions converge.
• Use canonical only for genuinely duplicate content, such as sort/filter parameter URLs pointing back to a base URL. Applying it too readily to variants that differ in price or stock — and that you want indexed independently — can cause the non-canonical pages to drop out of the index. canonical is a hint to Google, not a directive it is guaranteed to follow.
• For discontinued products, redirect to a related item or apply an appropriate status rather than leaving a dead page; for temporary out-of-stock items, setting availability to OutOfStock while keeping the page live is the common approach.

Images and first-hand information

Provide multiple angles, in-use shots, and size-comparison images, each with concise, descriptive alt text — not a mechanical list of the product name. Optimize file size and format (e.g. WebP) for load speed, since slow images cost pre-purchase conversions. Build differentiation with information only your company can produce: self-measured dimensions and weight, real hands-on notes and caveats, staff use-case recommendations, and questions and answers gathered from buyers — all written as text, not left only in images.

Structuring for AI, and Merchant Center

For ChatGPT, Perplexity, Google AI Overviews and similar systems to pick up product information, it needs to exist as readable HTML text — not only inside an image or rendered exclusively by JavaScript. Organize specs into tables, present features and use cases as headed lists, and phrase common questions as explicit question-and-answer pairs; leading with a direct, "answer-first" sentence makes the key point easier for AI to extract. Structured data reinforces the machine-readability of that text — it does not substitute for it. Registering with Google Merchant Center can make a product eligible for free listing (e.g. in the Shopping tab), a separate channel worth considering alongside product-page SEO, but registration does not guarantee listing or ranking — it still depends on meeting data-quality and policy requirements.

Measuring and prioritizing improvements

• Search Console — impressions, clicks, position, and query per product page, to see which search intents are actually surfacing the page.
• Structured-data errors — check the Merchant/Product enhancement report in Search Console regularly for invalid fields or warnings.
• Indexation — confirm important product pages are indexed, and that duplicate or parameterized URLs are not being over-indexed instead.

• 1. Remove compliance risk — fix any structured data that does not match what the page displays, including unshown review markup.
• 2. Clean up duplicate and thin pages — add first-hand information to manufacturer-copy-only pages; normalize unnecessary parameterized URLs.
• 3. Strengthen top products first — prioritize the products with the most traffic or stock, adding original information, FAQs, and images.
• 4. Implement structured data accurately — limited strictly to what is displayed, for Product/Offer.

Structured data (Schema) basics for BtoB sites（日本語）

Reading search intent: four categories and body-copy design（日本語）