Category and product-page optimisation: an implementation roadmap for small agencies

Start with the job each page type needs to do

Category and product-page optimisation is not one task. A category page helps people browse a useful set of products; a product page helps them evaluate one product or a family of variants. The SEO work is to make those roles clear, keep important pages discoverable through normal links, control duplicate URL combinations, and keep product data consistent with what a visitor can actually see.

Google says it generally infers ecommerce site structure from the links between pages rather than from the folder depth of the URL. In practical terms, an agency should first ask whether an important product can be reached through crawlable links from a logical category path—not whether the URL contains the perfect keyword sequence. See Google's guidance on ecommerce site structure and internal linking.

What category and product-page optimisation can actually change

Good implementation can improve discovery, reduce duplicate or low-value crawl paths, make product information easier for Google to interpret, and give users a clearer route from a broad category to the right item. It does not guarantee rankings, and valid structured data does not guarantee that Google will show a rich result.

• Discovery: Google recommends linking through the site hierarchy—for example, from category to subcategory to product—and notes that products reachable only through a site-search box may not be discovered through crawling alone.
• Crawl efficiency: faceted navigation can create extremely large URL spaces. Google's current guidance says unwanted facet URLs should be prevented from consuming crawl resources, while facets that genuinely need to be crawled should use a controlled, consistent URL design. For larger catalogues, the crawl-budget decision matrix for ecommerce sites gives a wider prioritisation framework.
• Search appearance: Product structured data can make eligible product pages suitable for richer product appearances, but Google recommends focusing that markup on pages for one product or variants of the same product—not category listings.
• Commercial accuracy: price, availability and variant information need to agree across the visible page, structured data and any Merchant Center product data. A technically valid annotation that contradicts the page is not a successful implementation.

Diagnostic checklist: identify the failure before editing

Run the diagnosis by template and by business importance. Start with a representative sample of high-value categories and products, then expand only when you know which failure pattern you are looking for.

For URL design, Google currently recommends stable, persistent URLs; descriptive words can help Google understand a page, but that is not a reason to rename working URLs solely to insert keywords. Google's ecommerce URL guidance also stresses consistency between internal links, canonicals and sitemaps.

Implementation roadmap: fix the highest-risk layer first

Sequence the work from access and discovery to product semantics and enhancements. This reduces the chance of polishing pages that Google cannot reliably discover, or adding markup to URLs you later decide should not exist.

Phase 1: capture a baseline and fix discovery blockers

1. Record the before state. Export the important category and product URLs, their current Google Search queries, clicks, impressions and click-through rate where available. Pair that with analytics data for organic sessions, enquiries, purchases or other conversions. Google recommends using Search Console and Analytics together because they answer different questions: Search Console covers performance before the visit; Analytics covers behaviour after arrival. See Google's guide to using Search Console and Google Analytics data together.
2. Make the product path crawlable. Link from relevant category pages to subcategories and products with normal anchors. If category content is paginated, make sure later pages have unique URLs and crawlable links rather than existing only behind a button or scroll event. Google's pagination guidance explains why this matters.
3. Do not rewrite stable URLs just for keywords. If the current URL is usable, canonicalised and already indexed, leave it alone unless there is a stronger architecture or platform reason to change it. When a URL change is genuinely necessary, redirect the old URL to the true replacement and update internal links, canonicals and sitemap entries to the final URL.
4. Resolve indexability and canonical conflicts. Check that the pages you want indexed return a successful response, are not blocked by an accidental noindex, and reference the intended canonical URL. Sample Google-selected canonicals in Search Console after deployment rather than assuming the declaration was accepted.

Phase 2: control facets, variants and inventory states

This phase is where ecommerce sites often create far more URLs than they intend. Treat each mechanism separately.

• Faceted navigation: decide which filtered URLs, if any, deserve to be searchable landing pages. If you do not need a facet URL indexed, Google's current crawling guidance recommends preventing unwanted facet crawling rather than leaving an effectively unlimited URL space open. If a facet does need to be crawled, use consistent parameters or paths, keep filter ordering predictable, prevent duplicate combinations, and return 404 for impossible/no-result combinations. See Google's faceted navigation guidance.
• Product variants: give a directly addressable URL to variants that Google needs to distinguish, and make sure that URL actually opens with the correct variant selected. Google's current ProductGroup guidance supports both single-page and multi-page variant designs, but the markup and canonical approach differ between them. Review the official product variant structured-data documentation before standardising a template.
• Temporary out of stock: do not return 404 simply because inventory is temporarily unavailable. Keep a useful product page live when the product is expected to return, state the availability clearly, and keep structured data and Merchant Center data consistent. Google's Merchant Center guidance explicitly supports out_of_stock for temporarily unavailable products.
• Permanently discontinued: remove the discontinued offer from Merchant Center product data. For the webpage itself, ScanMySEO recommends choosing deliberately: keep the page if it still helps users, redirect only when there is a genuinely equivalent successor, or return 404/410 when the product is gone and there is no useful replacement. Do not redirect every retired SKU to a category or homepage simply to avoid a 404.

For stock handling, see Google's product availability guidance, which requires the availability submitted to Google to match the landing page.

Phase 3: improve category content and product data without schema overreach

Once discovery and URL behaviour are stable, improve what the pages actually communicate.

• Category pages: give the page a clear category name, useful subcategory or filter choices, crawlable product links and enough explanatory content to help a buyer understand what belongs in the category. There is no useful universal word-count target; do not add paragraphs that merely repeat product-card text.
• Product pages: include the information a buyer needs to choose correctly. Depending on the product, that may include dimensions, materials, compatibility, technical specifications, minimum order quantity, lead time, installation constraints, variants, price, availability, shipping, returns, documentation or a clear enquiry path. For B2B catalogues, the practical decision information often matters more than generic sales copy.
• Use Product structured data on the right pages. Google's merchant-listing documentation says product rich results focus on a single product or variants of the same product, not a category containing unrelated products. Merchant-listing eligibility also applies to pages where a shopper can purchase the product; non-transactional product pages have different product-snippet requirements. Review the current merchant listing Product requirements before treating schema as a template-wide checkbox.
• Keep markup consistent with the page. Price, currency, availability, variant identity, ratings and other marked-up properties should reflect what visitors can see. If your ecommerce platform changes these values with JavaScript, test the rendered output and consider Google's recommendation to place Product markup in the initial HTML for merchant pages where practical.
• Use Merchant Center as a complementary data source when appropriate. Google recommends structured data on product pages and, for merchants, a Merchant Center feed for stronger control over product discovery and updates. The two sources can complement one another; one is not a substitute for accurate visible product information. See Google's guidance on sharing product data with Google.

Phase 4: turn recommendations into verifiable agency tickets

A small-agency roadmap is more useful when each recommendation has an owner and an acceptance test. Avoid tickets such as “improve category SEO” that cannot be checked after delivery.

Verification loop: prove the fix, then measure the outcome

Verification has two layers: first confirm that the implementation is technically correct, then measure whether user and search performance changed. Do not combine those into one vague “rankings improved” test.

1. Re-crawl the changed templates. Confirm status codes, indexability, canonical targets, internal links, pagination paths, facet behaviour and rendered product content. Compare a before-and-after sample so that a fix in one template has not created a regression elsewhere.
2. Validate Product markup. Use Google's Rich Results Test and the relevant Search Console product reports. A valid result means the markup meets the tested requirements; it does not mean Google will display a rich result. Google's merchant-listing guidance also recommends testing a small deployment and inspecting live URLs before rolling a change across a catalogue.
3. Inspect representative URLs in Search Console. Check the indexed state, Google-selected canonical and live rendering for important categories, products and variants—especially after canonical, JavaScript or URL changes.
4. Measure by page type and query intent. In Search Console, compare clicks, impressions, click-through rate and the queries reaching the changed pages. In Analytics, compare organic sessions and the business actions that matter, such as enquiries or purchases. Use matched periods where seasonality allows rather than attributing every short-term movement to the deployment.
5. Keep the conclusion proportional to the evidence. If crawlability is fixed, you can say the page is now easier to discover. If Product markup is valid, you can say it is eligible under the relevant rules. Only claim a traffic or conversion improvement when the performance data shows it.

Google's 2026 guidance for generative AI features does not create a separate category-page optimisation playbook: the same crawlability, indexing, useful content and ecommerce data fundamentals remain relevant, and Google says there is no special AI schema or required content “chunking”. That is a useful guardrail for small agencies deciding where limited implementation time should go. See Google's generative AI Search guidance.

The practical order is simple: make priority products discoverable, control duplicate URL spaces, handle variants and inventory truthfully, add product data only where it matches the page, and verify every release against a baseline. That gives a small agency a defensible implementation process instead of a checklist of unrelated “SEO fixes”.