Product page SEO: the complete guide

A product page wins specific searches and loses general ones, so aim it at the searches that name a thing rather than describe a category. Trying to rank a single product for a broad category term is the most common wasted effort in store SEO, because the result page for that term is full of pages offering choice and your page offers one option.

The searches a product page can realistically take are these. The exact product name, as printed on the box or as the manufacturer calls it. The model or part number, which people search when they are replacing something. The name plus a qualifier such as a size, colour or specification. The problem stated with the product implied, where the product is unusual enough to be the answer. And your brand name plus the product, if it is your own product.

For stores selling their own products, the situation is better and worse. Better because nobody else has your page, so the duplicate problem does not exist. Worse because nobody is searching for a product name they have never heard of, which means the demand has to be created elsewhere and the page's job is mostly to convert.

For resellers, the reality is that you are competing against every other stockist and often the manufacturer. The differentiators available to you are content nobody else has: your own photography, your own honest assessment, real customer reviews, answers to the questions your team actually gets, and details the manufacturer left out.

That leads to a practical decision about effort. Do not write unique content for every product in a catalogue of thousands. Rank your products by revenue and by whether anyone searches for them by name, and give real attention to the top group. For the long tail, a consistent template with accurate specifications is enough.

Before you start on any page, search for the product name yourself and look at what ranks. If the results are all marketplaces and the manufacturer, be honest about what winning looks like, and consider whether a collection or a guide is a better route to that customer.

Pick one primary search per product page and a small set of related ones, then use the language customers actually use rather than the language on the invoice. Product pages fail more often from targeting nothing in particular than from targeting the wrong thing.

Start with the name. Manufacturers, distributors and customers frequently call the same item three different things, and the version that gets searched is usually the customer's. Check what people type by looking at your own on-site search logs, the queries your search console reports for the page, and the wording in your customer emails and reviews.

Collect the variations that matter. Abbreviations and full forms. British and American spellings if you sell across both. The generic category name people use when they do not know the brand. The compatibility phrasing, if your product fits something else, since a large amount of replacement-part demand is phrased that way.

Model and part numbers deserve their own attention. People searching a part number are close to buying and there is very little ambiguity in what they want. Make sure the number appears on the page in plain text, not only inside an image or a specification widget loaded by script.

Then decide the qualifier the page should carry. For a product available in several sizes or colours, the page usually targets the base product, with the variants handled inside it. Where a variant has its own distinct demand, such as a specific model rather than a size, it may deserve its own page.

Write these choices down against the product. A short note recording the primary search, the variations and any decision about variants keeps the page from drifting each time somebody edits it, and makes it possible for another person to work on it later.

One discipline that pays: never target the same primary search from two product pages. When two of your pages compete for one query, both perform worse than one would.

Write the page title so it reads as the answer to the search, with the most important words at the front. The title is what appears in results, so it is doing two jobs at once: telling a search engine what the page is about, and persuading a person to choose it over the results around it.

The structure that works for most products is brand, product name, then the distinguishing attribute. That order puts the words people searched for at the start, where they carry the most weight and where they are visible even when the title is trimmed. Adding your store name at the end is fine, and check whether your platform is doing it automatically so it does not happen twice.

Keep it honest and specific. A title padded with words like best, cheap or online adds nothing and takes space from something useful, such as a size, a quantity, a material or a compatibility note that helps someone recognise their product.

The visible heading on the page can differ from the title tag, and often should. The title has to work out of context in a list of results. The heading sits above the images and the price, where the shopper already knows roughly where they are. Use the heading for the clearest customer-facing name, and use the title to include the variations people search with.

The meta description does not affect ranking directly and it affects whether people click. Write it as one or two plain sentences saying what the product is and giving the reason to choose your store: delivery, returns, range, expertise. Avoid writing it as a list of keywords, which reads badly to the only audience that sees it.

One heading per page at the top level, describing the product. Do not use the top-level heading for a promotional banner or a section name, which some themes do by default and which muddies what the page is about.

Finally, check what your titles actually look like in results, not in the admin, because platforms and themes both modify them.

Write the description from what you know about selling the product, because the manufacturer's copy is on every other stockist's page and adds nothing to yours. This is the single largest differentiator available on a reseller product page.

Start with the first paragraph, which is what people read and what tends to get quoted in generated answers. Say what the product is, who it suits, and the one thing that makes it the right choice, in plain language. Do not open with a brand story or a sentence that could describe any product in the category.

Then write the things only a seller knows. Which customers buy this and why. What they usually pair it with. Where it sits against the alternatives in your own range, honestly, including who should buy the cheaper one. What gets returned and why, and how to avoid that. How it performs after a few months. Which situations it is not right for.

That last point does more for conversion than any amount of enthusiasm. A page that says clearly who a product is not for is more believable about who it is for, and it reduces returns.

Keep specifications separate from prose and make them complete. Measurements with units, materials, weight, what is in the box, power or compatibility requirements, care instructions. This is the information that gets pulled into generated answers and comparison features, and it is checkable, which is exactly why it is valuable.

Structure it so a scanning reader gets the point. Short paragraphs, clear subheadings for each area, the specification block in a consistent place across the catalogue.

Write for the person, not for a word count. A short description that answers everything is better than a long one that repeats itself, and padding is obvious to both readers and search engines.

For the long tail of products nobody searches by name, a consistent template with accurate specifications is a reasonable answer. Save the writing for the pages where it changes the outcome.

Put the answers to the buying questions on the product page itself, because a question left unanswered is a visitor who leaves to look elsewhere and does not come back. Your customer service inbox already contains the list.

Gather the questions properly. Ask whoever answers customer emails which questions come up every week. Look at pre-purchase chat logs. Read your reviews for the things people say they were unsure about. Look at the questions asked under your social posts. That list is the specification for this part of the page.

The recurring ones across most categories are delivery cost and timing, returns terms in plain words, sizing or fit, compatibility, what is included, how long it lasts, care and maintenance, and whether it does the specific thing the buyer needs.

Place them where the question arises rather than in a block at the bottom. Delivery and returns belong near the price, because that is where the concern occurs. Sizing belongs beside the size selector. Compatibility belongs near the specification. A frequently asked questions section at the foot of the page is better than nothing and worse than answering in context.

Write the answers plainly and specifically. A returns policy that says we accept returns is not an answer. One that says how long, in what condition, who pays the postage and how the refund arrives is.

This material also works hard for search. Question-shaped content matches question-shaped searches, and it is the kind of specific, checkable text that generated answers draw from. A page that clearly answers whether a product fits a particular situation can be surfaced for that exact query.

Make sure the answers are in the page's HTML rather than only appearing after a click or a script. Content hidden inside accordions is generally fine, and content that only loads after an interaction may not be read at all. Check the page source if you are unsure.

Keep them current. An answer about delivery times that was true two years ago is now a complaint waiting to happen.

Shoot your own product images, because they do more for both ranking and conversion than any other single change on a reseller product page. Supplier images are identical everywhere and they tell a shopper nothing about whether you are a real store.

The set that works covers a predictable list. A clean shot on a plain background, which is also what shopping feeds need. The product in use, at real scale, so people understand the size. Detail shots of the parts people ask about: the fastening, the seam, the connector, the texture. Anything that reveals quality. And a shot showing what actually arrives, including packaging, which reduces the surprise that causes returns.

Video does the job that photographs cannot. A short clip showing the product working, being fitted, or being unpacked answers questions no still image can, and it keeps people on the page. It does not need production values. It needs to show the thing.

Name the files descriptively before upload and write alt text that describes the image for someone who cannot see it. Both are quick, both help image search, and alt text is an accessibility requirement rather than an optimisation trick, so write it truthfully rather than stuffing it.

Watch the file sizes. Product images are usually the heaviest thing on the page, and a page that takes seconds to show its main image loses buyers who arrived ready. Serve appropriately sized versions rather than full-resolution originals scaled down in the browser, and make sure the main image is prioritised over everything below it.

Avoid text, badges and promotional overlays burned into the main image. They date badly, they read poorly at small sizes, and they cause problems in shopping feeds that use the same files.

Finally, keep the images consistent across the catalogue. A category page where every product was photographed differently looks disorganised, and consistency in background, angle and crop makes a store look like a store.

Collect reviews systematically and display them on the product page, because they add unique content, answer objections and change whether people buy. They are the one form of page content that gets better on its own once the process exists.

Set up the collection properly. An automated request after delivery, timed for when the customer has actually used the product rather than when it shipped, sent once with a single polite reminder. Make leaving a review take under a minute. Ask a specific question rather than requesting general feedback, because a prompt like what were you unsure about before ordering produces reviews that are useful to the next buyer.

Display them on the page with the reviewer's context where possible, since a review from someone with the same use case is worth more than a generic five-star line. Show the negative ones. A page with only perfect reviews reads as filtered, and a well-handled criticism with a reply from the store is more persuasive than another compliment.

Answer reviews that raise a question or a problem. That reply is read by everyone considering the product, and it demonstrates what happens when something goes wrong.

On markup, review and rating data can be included in your product structured data so search engines can display it, and only when the reviews are genuine, from customers, and actually visible on the page. Marking up ratings that are not shown, or that were not left by buyers, creates a problem that is worse than the gap it fills. Check that you do not end up with two sources of review markup, since review apps and themes often both add it.

For new stores with no reviews, be patient rather than inventive. Ask early customers directly, offer to replace anything they are unhappy with, and accept that the first ones take effort. Fabricated reviews are visible to customers, damaging to trust, and a genuine legal risk in many markets.

Mine the reviews for everything else too. The words customers use are the best source of copy you will find.

Decide deliberately how variants and stock states behave, because the defaults create duplicate addresses and dead pages that nobody notices for months.

On variants, the general rule is one page for the product with the options selectable inside it, and every variant address resolving back to that page. This concentrates the value in one place and matches how people search, since most people search for the product rather than for a specific size.

The exception is a variant with genuine independent demand: a distinct model, a materially different specification, or something people search for by its own name. Those earn their own page. The test is whether anyone searches for that variant specifically, not whether it feels different internally.

Make sure the page content updates properly when a variant is selected. Price, availability, images and the specification should all reflect the selection, and the key facts should still be present in the page source rather than only appearing after a script runs.

On stock, a temporarily unavailable product should keep its page, say clearly that it is out of stock, give a return date if you have one, and offer both alternatives and a notification option. Hiding it removes internal links and loses the position. Redirecting it sends people somewhere they did not ask for.

For products that are gone permanently, redirect to the closest equivalent. Where there is a natural successor, that is the destination. Where there is not, the category is better than the home page. Only leave a page live with no product when the demand for it is worth serving with alternatives.

Near-duplicate products deserve a decision too. Catalogues accumulate items that are almost the same, listed separately for historical reasons, and they compete with each other in search. Merging them into one page with options is usually better for both ranking and shoppers.

Audit this once a year. Every catalogue drifts, and an hour spent on stock states and duplicates recovers pages nobody realised had gone quiet.

Link product pages properly from the rest of the site, because a page nobody links to is crawled rarely and ranked poorly no matter how good it is. Products sit at the bottom of most site structures and receive the least internal attention by default.

Start with the categories. A product should be linked from every collection it genuinely belongs in, which is often more than the one it was filed under. Products that exist in only one deep category are the ones that go unnoticed.

Add breadcrumbs. They give a shopper their position, they create a link back to the parent category from every product, and they express a hierarchy that many store URLs do not. On platforms with flat URLs, breadcrumbs are doing more work than they appear to.

Use related products with judgement. Automated modules generate a lot of links with no editorial thought, which dilutes the signal from the links you chose. Where you can, curate them: what people actually buy together, the upgrade, the alternative for a different use.

Link from your supporting content. A sizing guide, a buying guide or a comparison article should link to the specific products it discusses, with text that names them. These links carry more weight than template-generated ones because they were chosen, and they help the reader.

Check which of your product pages receive almost no internal links at all. A crawl that lists pages by inbound internal link count usually reveals a large group with one or two, and those are the pages that never rank. Fixing that is often just a matter of adding them to the right collections.

Finally, make sure product links across the site point at the current address rather than at a redirect. Old links accumulate after renames and migrations, and each one adds a hop that costs a little and looks untidy in a crawl.

Judge a product page on the orders it produces from the searches you aimed it at, and stop working on it when the remaining gains are smaller than the same effort spent elsewhere. Store SEO wastes a lot of time polishing pages that were already fine.

The measures worth watching are simple. Whether the page ranks for the primary search you chose. How many organic sessions it receives. What share of those visitors add to cart. What share complete an order. And the revenue, which is the only one that pays.

Read them together. A page with impressions and no clicks has a title or a listing problem. A page with clicks and no add to cart has a page problem: unanswered questions, a price surprise, weak images or a missing detail. A page with add to cart and no orders has a checkout or a delivery cost problem, which is usually not the page's fault at all.

Watch the queries the page actually attracts. If a page is receiving searches you did not aim for, that often reveals demand worth serving properly, either by adjusting this page or by creating a better one. If it receives nothing, the words were wrong.

Be realistic about timescale. A new product page takes time to settle, and edits made every week make it impossible to know what worked. Change something, wait for enough data, then judge.

Work in batches by value. Take the products that produce the most revenue or have the most search demand, improve them properly, and move on. Then take the group of pages ranking just outside the top results, since a small improvement there produces a large traffic change.

A page is finished when it answers the buying questions, has your own images, carries genuine reviews, loads quickly and ranks where the search results allow. At that point, more editing is decoration. The next gain is on a different page.