product seo

Product pages are the strangest asset in organic search. They are the closest page to revenue and the most likely to be neglected, usually because they are generated by a platform template that nobody has looked at since launch. Last updated August 28, 2026, this is what I have learned running product SEO across catalogues ranging from 40 SKUs to 1.2 million, and the specific interventions that moved money rather than rankings.

I work almost exclusively on ecommerce, and I want to lead with the most important framing I can offer. Product SEO is not a content exercise. It is an inventory, architecture and data quality exercise that happens to involve writing. The clients who understood that improved fastest.

Why Most Product Pages Fail Before A Word Is Written

The typical failure looks like this. A store has 8,000 products. Each product page carries the manufacturer's supplied description, which is identical to the description on 200 other retailers' sites. The title tag is the product name plus the store name. There are three photographs from the supplier. There are no reviews. The page has two internal links, both from paginated category listings that are themselves noindexed.

That page cannot win, and no amount of rewriting will fix it while the structural conditions hold. On one home goods client I audited in 2024, 71 percent of product URLs had zero internal links from any non-paginated source. They had spent 40,000 dollars on product descriptions the previous year and seen no movement, which was entirely predictable.

The Product Page Template That Actually Performs

Across my accounts, the pages that rank and convert share a consistent anatomy. I will describe it concretely because vague advice is useless here.

The title tag leads with the product's distinguishing attribute set rather than the bare name. For a specialist retailer selling a hiking boot, "Waterproof Leather Hiking Boot, Wide Fit, Vibram Sole" beats the SKU name, because the SKU name is what everyone else uses and the attributes are what people search.

The H1 stays as the human product name so the page does not feel like a keyword exercise. Immediately below it sits a two-to-three sentence summary that answers who the product is for and what makes it different. On a fashion client, adding that summary block above the fold raised add-to-cart rate 11 percent, which had nothing to do with SEO and everything to do with the fact that nobody had ever told the shopper why to choose it.

Then unique descriptive copy of 150 to 400 words, written by someone who has handled the product. Then specifications in a structured table, which is the section that actually gets extracted by answer engines. Then reviews. Then a small set of genuinely relevant cross-links.

Photography deserves its own note. Sites that replaced supplier images with their own shots consistently outperformed, and my working theory is duplication plus trust rather than any image ranking factor. On a kitchenware client we reshot 300 top products and that cohort grew organic revenue 34 percent against a control group at plus 6 percent over the same five months.

Structured Data Is Not Optional For Products

Product schema is the highest-return technical work in this entire discipline, and it is astonishing how often it is wrong. I audit it on every engagement and I find errors in roughly seven out of ten stores.

The common failures are specific. Price and currency mismatched between schema and the visible page. Availability hardcoded to InStock regardless of actual inventory. Aggregate rating markup present on products with no reviews, which is a policy violation. Missing GTIN or MPN on products that have them. Review markup nested incorrectly so it validates but does not produce enhancements.

Fixing schema on a 4,000-product electronics client produced rich result eligibility on 3,100 URLs within three weeks and a 19 percent click-through rate increase on the affected pages. No content changed. That is the cheapest win available in product SEO and it requires a developer for a week, which is why platform-generated schema needs auditing rather than trusting.

Faceted Navigation: The Silent Catalogue Killer

Every large store I have worked on has had a faceted navigation problem, and it is the issue that separates product SEO from general SEO.

The mechanics are simple. Filters generate URL parameters. Parameters generate URLs. If those URLs are crawlable and indexable, a 5,000-product store can generate 400,000 near-duplicate pages. Crawl budget evaporates, and Google spends its time fetching "blue medium cotton sorted by price ascending" instead of your actual products.

My standard approach: choose a small set of facet combinations that have genuine search demand, usually attribute plus category, and make those crawlable, indexable, and given real content and titles. Everything else gets canonicalised to the parent category and blocked from crawling at the parameter level. Sorting and pagination parameters never become indexable, ever.

On a fashion retailer with 12,000 products, we cut indexable URLs from 380,000 to 19,000. Crawl requests to product pages tripled within a month. Organic revenue rose 41 percent over two quarters. Nobody wrote a single blog post. This kind of change needs real engineering discipline, and it is worth having it reviewed by whoever handles your web applications rather than configured through a plugin.

Handling Out Of Stock And Discontinued Products

This is where I see the most avoidable revenue loss, because the instinct is to delete and deleting destroys accumulated authority.

My rules, developed after getting this wrong on a seasonal retailer in 2019 and losing 30 percent of their autumn traffic: if the product is temporarily out of stock, keep the page live, keep it indexed, mark availability correctly in schema, and add a restock notification capture. If the product is permanently discontinued but has a successor, keep the page live with clear messaging and a prominent link to the successor, then 301 it only once traffic has decayed. If it is discontinued with no successor, 301 to the most specific relevant category, never to the homepage.

On the seasonal retailer, switching from delete-and-redirect to keep-and-notify recovered the traffic the following year and generated 4,200 email subscribers from restock forms, which turned an SEO fix into an email marketing asset.

Reviews Are A Ranking And Conversion System

User-generated content on product pages solves two problems simultaneously. It gives the page unique text that no competitor has, and it gives the shopper the social proof that closes the sale.

The numbers from my accounts are consistent. Products with more than twenty reviews convert at roughly 1.6 to 2.2 times the rate of products with fewer than five, controlling for price band. And review text surfaces long-tail queries you would never have researched, because customers describe products in language your copywriters do not use. I mine review text for keyword ideas on every account and it outperforms tool-based research for product attributes.

Category Pages Do The Heavy Lifting

An uncomfortable truth about product SEO: for most head and mid-tail commercial queries, your category page is the asset that ranks, not your product page. Product pages win on model numbers, exact names and highly specific attribute strings.

That reframes the work. Category pages need real introductory content, buying guidance, curated subcategory links and internal links from your informational content. On a sporting goods client, adding 600 words of genuine buying guidance plus a comparison table to eighteen category pages produced an average gain of 4.8 positions and a 27 percent revenue increase for those categories over four months.

The mistake to avoid is dumping 1,500 words of keyword-stuffed text below the product grid where nobody reads it. Put guidance where it helps the shopper, keep it short at the top and expandable, and write it as though a salesperson wrote it. Getting this right is a design problem as much as a copy problem, and it is usually the point where a front-end web development conversation becomes necessary.

Speed On Product Pages Is Revenue, Measurably

Product pages carry the heaviest payloads on any store: image galleries, review widgets, recommendation carousels, personalisation scripts, chat, analytics, and usually three tag manager containers nobody has audited.

On a supplements retailer we cut mobile largest contentful paint from 5.4 seconds to 2.1 by lazy-loading gallery images beyond the first, deferring the review widget until scroll, and removing four abandoned marketing tags. Conversion rate on mobile rose 16 percent within a fortnight. Organic rankings moved modestly. The revenue impact came from conversion, and that is the honest framing for speed work on product pages.

How I Prioritise Work Across A Large Catalogue

With a million SKUs you cannot optimise everything, so prioritisation is the actual skill. I build one spreadsheet per engagement with five columns per product: 90-day organic impressions, 90-day organic revenue, gross margin, internal link count, and whether the description is unique.

Products with high impressions and low clicks go first, because the demand already exists and something on the page is losing the click. Products with high margin and zero impressions go second, because they are invisible and worth being visible. Everything else waits.

That ordering has produced faster revenue movement on every catalogue I have applied it to than any topic-cluster approach, because it starts from demonstrated demand instead of hypothesised demand. On a pet supplies client, the first twelve weeks of work touched 340 SKUs out of 61,000 and delivered a 22 percent organic revenue increase, and the reason was simply that we worked the list in the right order.

The other prioritisation call worth making early is which products deserve supporting content and imagery. Complex or considered purchases benefit enormously from comparison charts and explanatory diagrams, and commissioning that properly through a graphic design resource is usually cheaper per revenue point than another thousand words of description.

Frequently Asked Questions

Closing Thoughts

The best product SEO work I have done looks nothing like content marketing. It looks like fixing a schema block, blocking 360,000 junk URLs, reshooting photographs, and getting review requests sent five days after delivery instead of at dispatch. None of it is impressive in a slide deck and all of it showed up in revenue.

If you have a catalogue and limited resources, start with an inventory of what is indexable and what is internally linked. That single spreadsheet has told me where the money was hiding on every ecommerce engagement I have taken, without exception.