Most store owners get seo for ecommerce product pages backwards. They pour effort into individual SKU pages, paste in whatever the manufacturer sent, and wonder why a catalog of 4,000 products pulls 300 organic visits a month. The pages exist. They just have nothing to say that a hundred other retailers aren’t already saying with the same words.
Product pages can carry serious revenue. They just win a different kind of query than most people assume, and the work that makes them win is unglamorous: deciding which ones deserve attention, writing copy nobody else has, marking it up correctly, and having a plan for what happens when a product dies.
Product pages win long-tail, not head terms
Here is the split that governs everything else. Category pages usually carry more commercial search volume than product pages. “Running shoes,” “standing desks,” “wireless earbuds” — those are category-level demand, and a well-built category page is the right landing target. Product pages win the long tail: exact brand-plus-model queries, part numbers, “Brand X Model 400 review,” “Model 400 vs Model 500,” and the specific spec-qualified searches that convert at two to four times the rate of head terms.
That changes your prioritization. A product page fighting for a generic two-word term is usually losing to a category page — often your own. Check it: if your product URL and your category URL both target the same phrase, you have keyword cannibalization, and Google will pick one, usually the weaker one. Assign the head term to the category, the model-specific terms to the product.
The practical test is search intent. Type the query into Google. If page one is mostly category and collection listings, that term belongs to your category page. If page one is individual product listings, reviews, and manuals, the product page is the right target.
The manufacturer description trap
This is the single most common cause of thin product pages, and it’s structural rather than lazy. The manufacturer ships a description. Forty retailers paste the same 90 words. Google now has forty near-identical documents and has to pick a winner, which will be whichever site has the strongest signals — usually Amazon or the brand’s own store. Everyone else gets filtered as duplicate.
You do not need a thousand words per SKU to escape that. You need something original that the source text doesn’t contain: who the product is actually for, what it does not do well, how it compares to the two nearest alternatives in your own catalog, sizing or fit notes from returns data, compatibility warnings your support team keeps repeating. Two original paragraphs of that beats 600 words of rewritten spec sheet.
Keep the manufacturer copy if you want — customers sometimes want the official spec language. Put it below your original content, not above it. And never let it be the only prose on the page.
What belongs on a product page
Structure matters more than volume. A page that answers buying questions in a predictable order outperforms a wall of text, because real shoppers scan and Google is increasingly good at reading structure.
- Title tag: brand + model + key qualifier + your differentiator. “Acme Trail 400 Hiking Boot — Wide Fit, Free Returns” beats “Acme Trail 400 | Store Name | Buy Online.”
- H1: the product name as a customer would say it, not the internal SKU string.
- Original opening: 60–120 words that position the product against alternatives.
- Spec table: real attributes in a table, not buried in prose. This is what feeds comparison queries.
- FAQ block: three to six genuine pre-purchase questions pulled from your support tickets and live chat logs.
- Reviews: user-generated content is the cheapest source of unique text at scale, and it updates itself.
- Internal context: related products, accessories, “customers also viewed” — real crawl paths, not just merchandising.
One caution on images: they sell the product but they don’t rank the page. Compress them, give them descriptive alt text, and lazy-load below the fold. A 4MB hero image on mobile costs you more in Core Web Vitals than the alt text gains you.
Product schema, and where it goes wrong
Product structured data drives price, availability, and review rich results in the SERP. Those snippets don’t directly move rankings, but they move click-through rate, and on a competitive listing page a star rating and an in-stock price can be worth a full position of traffic.
Mark up the essentials: name, image, description, brand, SKU or GTIN, offers with price and priceCurrency and availability. Keep availability accurate — schema that claims InStock on a sold-out page is a fast route to a manual action against your rich results.
Review markup has a specific rule people ignore: it must not be self-serving. You cannot mark up reviews you wrote about your own products. Aggregate ratings from genuine customers on the product page are fine. Ratings you invented, or ratings for the store itself embedded in Product markup, are not, and Google has enforced against both.
Out-of-stock and discontinued products need a policy
Silent deletion is the default in most ecommerce platforms and it’s the worst option. A product page that ranked for two years and accumulated links becomes a 404, the link equity evaporates, and any customer who bookmarked it hits a dead end.
Pick a rule per situation and apply it consistently:
- Temporarily out of stock: keep the page live, keep it indexed, mark availability as OutOfStock in schema, and surface a restock date plus related in-stock items. Do not redirect. The page still answers the query.
- Permanently discontinued, with a clear successor: 301 to the nearest equivalent product. The Model 400 redirects to the Model 500. Redirecting to the homepage or a top-level category is treated as a soft 404.
- Discontinued, no equivalent, no traffic, no links: return 410 Gone and let it drop cleanly. This is the right call for the long tail of dead SKUs clogging your crawl budget.
- Seasonal products: keep the URL permanently. Never delete and recreate a Christmas category every year — you throw away its history each time.
You cannot hand-write 40,000 pages, so prioritize
Every large catalog runs into the same wall. The honest answer is that most SKUs will never justify bespoke copy, and pretending otherwise stalls the whole project.
Rank your catalog by opportunity before writing anything. Pull search volume for brand-plus-model terms across your product set, cross-reference against margin and stock depth, and you’ll typically find that a few hundred products account for the great majority of realistic organic upside. Write those properly. For the rest, use structured templates that pull real attribute data — genuinely different specs produce genuinely different pages — plus reviews and Q&A to accumulate unique text over time.
This is where tooling earns its keep. SEO Rocket’s keyword research gives you volume, difficulty, and country-specific data on model-level terms so you can decide which SKUs are worth the effort, its content gap tool shows what competing retailers rank for that you don’t, and its AI writer produces category and product copy that isn’t recycled manufacturer boilerplate. Full-site crawls have been verified at 900+ pages and surface duplicate titles, thin pages, and crawl problems at catalog scale. It does not manage product feeds, Merchant Center, or inventory — that stays in your ecommerce platform.
Measure trends, not daily positions
Product page work is slow to show results. New pages typically need four to twelve weeks before positions stabilize, and daily movement of two or three positions is ordinary noise, not a signal. Judge by four-week trend lines, and treat third-party volume and position data as modeled estimates rather than fact.
Your own Search Console data is the ground truth that matters. Watch impressions before clicks: a product page gaining impressions on model-specific queries is working even before it cracks the top ten. If impressions are flat after eight weeks, the problem is usually the page, not patience — and nine times out of ten, it’s still the manufacturer description sitting where your original copy should be.
