Ecommerce SEO: fix the system before writing more product copy

When an online shop needs more organic revenue, “write more product descriptions” is an easy recommendation. It is often the wrong first move. If search engines cannot reach the important pages, filters create thousands of duplicates, or product data is inconsistent, more copy only decorates a broken system.

I approach ecommerce SEO as product work: understand demand, make the catalogue crawlable, give each useful page a clear job, and measure what happens after the click. Rankings are an input. Revenue, margin and qualified discovery are the business outcomes.

Start with demand and the catalogue, not a generic checklist

Before touching templates, map search demand to page types. Broad commercial queries usually belong to categories; specific model or attribute queries may belong to products or indexable filtered pages; comparison and how-to questions belong to guides.

Then compare that map with the catalogue and analytics:

• Which categories and products already receive impressions in Search Console?
• Which pages generate organic transactions and margin?
• Where does stock stay stable enough to justify long-term optimisation?
• Which search intents have no useful landing page?
• Which URLs compete for the same intent?

Do not interpret “focus on 20% of products” as a universal Pareto law. Use actual revenue, margin, availability and search opportunity to create a manageable priority set. The point is sequencing, not a magic percentage.

Make navigation and crawl paths explicit

Google primarily discovers ecommerce pages through links. Important products should be reachable through normal <a> links from categories, subcategories and relevant editorial content. A search box or JavaScript-only interaction is not a reliable substitute for crawlable navigation.

Keep the hierarchy understandable: home, category, subcategory when necessary, product. Breadcrumbs help people retrace that structure and can reinforce it for search engines when paired with valid structured data.

Faceted navigation needs a deliberate policy. Some combinations represent real demand and deserve stable landing pages. Most combinations — every size, sort order and price range — do not. Decide which facets may be indexed, keep internal links consistent, and prevent crawl traps. A canonical tag is a hint for duplicate consolidation, not a guarantee that unlimited parameter URLs will stop being crawled.

The XML sitemap should contain canonical, indexable URLs you want in search. It should not be an inventory dump containing redirects, blocked pages and alternate parameter versions. The free sitemap inspector and robots.txt generator can help with a first pass, but Search Console and server logs show what Google actually does.

Give categories and products different jobs

A category page should help someone narrow a choice. Its title, heading, introduction, filters, product ordering and internal links should reflect a coherent buying intent. There is no required word count. A short explanation that clarifies range, compatibility or delivery is more useful than a long block written to hit keywords.

A product page should resolve the questions that stand between interest and purchase:

• accurate name, price, availability and variant information;
• original specifications and compatibility details;
• clear images with useful alternative text;
• delivery, returns and warranty information;
• reviews and questions when they are genuine and moderated;
• a stable canonical URL and sensible behaviour for unavailable items.

Manufacturer copy repeated across many retailers gives neither customers nor search engines much new information. Original content matters when it adds evidence: measurements, fit notes, usage guidance, comparisons, photography or support answers.

For discontinued products, choose the response by user value. Keep a page live when it still answers useful questions or has a clear replacement. Redirect only to a genuinely equivalent destination. When nothing useful remains, a 404 or 410 is more honest than sending every old product to the homepage.

Treat product variants and structured data as data modelling

Variants do not have one universal URL pattern. A colour may deserve its own URL when people search for it and it has distinct images or availability; a size often does not. Whatever model you choose, keep URLs, canonical signals, internal links and structured data consistent.

Google supports product structured data and merchant listings with properties such as offers, price, currency and availability. Markup must describe what is visibly present on the page. It can make a page eligible for richer presentation, but it does not guarantee a rich result or ranking improvement.

Validate representative templates, then monitor errors at scale. The schema generator can help inspect the shape of a block; Google’s Rich Results Test and Search Console are the final checks for Google eligibility.

Performance and UX are part of the same system

Core Web Vitals are used in Google’s page experience systems, but good scores alone do not make a page relevant. Their more immediate value is practical: fast, stable and responsive product pages are easier to use.

On ecommerce sites I usually inspect:

• whether the main product image is discoverable early and properly sized;
• image formats, responsive sources and cache policy;
• layout shifts from galleries, banners and review widgets;
• long interaction delays caused by variant selectors or third-party scripts;
• server response time for uncached category and search pages.

Do not apply lazy loading to the likely LCP image. Remove or defer tags that do not justify their cost, and judge real-user data separately for key templates. A laboratory score is a diagnostic snapshot, not proof of customer experience.

Measure SEO as a path to revenue

Search Console records search visibility and clicks; GA4 and the commerce backend record behaviour and orders. None is a perfect source on its own. Join them at the level your data quality supports, and keep attribution limits visible.

A useful monthly view includes:

• organic clicks and impressions for priority page groups;
• index coverage and crawl behaviour;
• organic product views, add-to-carts, checkouts and purchases;
• revenue, margin where available, average order value and refunds;
• conversion rate by device and landing-page type;
• releases, stock changes, promotions and seasonality that explain the trend.

Avoid promising a specific position or a fixed month for results. Crawling, competition, site history, implementation speed and demand all matter. Technical fixes can change coverage quickly, while competitive category growth may take much longer. Set review points around shipped work and leading indicators, not a generic 3–6–18 month script.

A practical order of work

1. Connect Search Console, analytics, backend revenue and crawl evidence.
2. Map demand to categories, products, useful facets and guides.
3. Fix discovery, indexing, canonicalisation and status-code mistakes.
4. Improve the highest-opportunity category and product templates.
5. Make product data and structured data accurate and consistent.
6. Address real performance and accessibility bottlenecks.
7. Build useful editorial assets and earn relevant references over time.
8. Measure by page group, learn, and expand the winning pattern.

Link earning works best when the destination stays useful: original research, buyers’ guides, tools, compatibility references and durable categories. Bought directory volume is not a substitute for relevance. The goal is a catalogue that people can understand and other sites have a reason to cite.

My role is usually the technical and interface layer: architecture, templates, performance, metadata, structured data and measurement. A specialist editorial or outreach team may own the wider content and authority programme. If the bottleneck is implementation, see my services, look through selected work, or describe the problem.

Sources

• Google Search Central: ecommerce site structure
• Google Search Central: faceted navigation
• Google Search Central: canonical URLs
• Google Search Central: ecommerce structured data
• Google Search Central: product structured data
• Google Search Central: Core Web Vitals