Ecommerce SEO: the complete guide

Search results for buying queries are crowded with things that are not stores, and understanding the shape of the page you are trying to win decides where you spend effort. Type a query your product would answer and look at what is actually there.

You will usually find some mix of the following: paid product listings across the top, a block of shopping results, a large marketplace or two, a review or comparison article from a publisher, a video, a generated answer summarising the topic, and somewhere below all of that, retailers. The share of the page available to a store like yours is smaller than the ranking position suggests.

This matters because it changes which queries are worth pursuing. A query where the results are dominated by publisher comparison articles is telling you that the search engine believes people want advice, not a shop. You can compete there, but you have to compete with a page that gives advice, not with a category listing of your products. A query where retailers rank is one where a well-built collection page has a real chance.

The second thing to understand is that your product pages are, from a search engine's point of view, mostly duplicates. If you resell products other stores also sell, the manufacturer's description is on hundreds of sites and yours has nothing to distinguish it. The pages that win in that situation are the ones that add something: real photography, honest detail, answers to the questions buyers actually ask, reviews from people who bought.

The third is that your own brand searches are the easiest traffic you will ever earn and the easiest to lose. Anyone searching your store name should find you, your key categories and your customer service pages immediately. Check this before anything else, because it is quick and occasionally broken.

Finally, be realistic about timescale. This channel compounds and it does not start quickly. Work started now pays over quarters, which is a reason to start rather than a reason to wait.

Start from your products and work outwards, because the fastest wins are searches you already have inventory to answer. Most keyword research for stores goes wrong by starting with a tool and a big list rather than with the catalogue.

Take your product range and write down every way a customer describes it. The category name. The specific use. The problem it solves. The material, size, colour or specification that people search by. The occasion or season. The compatibility, if it fits something else. Do this in the customer's language, which is frequently not the language on your website, and get it from customer service emails, on-site search logs and the questions in your reviews.

Your own site search is the most underused source here. It tells you exactly what people who reached your store expected to find, in their own words, and it flags the products they wanted that you do not have.

Then group those terms by what the searcher wants, because that decides which page type answers them. Some searches want a list of options to choose between. Some want one specific product. Some want to understand a difference before choosing. Some want to know whether something will work for their situation. Each of these needs a different page, and matching them wrongly is the most common reason a page never ranks.

Size the opportunity honestly. Volume matters less than fit. A search with modest volume that describes exactly what you sell is worth more than a high-volume category term you have no chance of winning, and it converts far better.

Then prioritise by three things together: how much you could sell if you ranked, how hard the result page looks to break into, and whether you already have a page that is close. Pages already ranking just outside the top results are the cheapest wins in SEO, because the hard part is done.

Write the map down as a table of searches, the page that should answer each, and whether that page exists. That table is your work queue for the next year.

Build the site so any product is reachable in a few clicks from the home page, through categories that match how people search. Structure is the part of SEO with the largest effect and the least attention, because it feels like a merchandising decision rather than a search one.

The principle is that your category structure should mirror your demand map. If people search for a type of product, there should be a category page for it. If people search for a specific attribute across categories, such as a material, a size or a use, there should be a page for that too. A store whose navigation reflects internal departments rather than customer language leaves those searches unanswered.

Keep the hierarchy shallow. Deep nesting buries products, dilutes the internal links reaching them, and makes crawling slower. Broad and shallow beats narrow and deep for almost every catalogue.

Category pages should sit on stable, readable URLs that describe the content, and those URLs should not change. When they must change, redirect properly, because a category page that has earned rankings over years loses them in a week if it moves without one.

Filtering is the structural decision that goes wrong most often. Filters create combinations, and combinations create pages, and a large catalogue can generate an enormous number of near-identical URLs that consume crawl attention and compete with each other. The approach that works is to decide which filter combinations represent real demand, give those a proper indexable page with its own text, and prevent the rest from being indexed at all.

Handle products that live in more than one category deliberately. Pick one canonical location and be consistent, so the same product does not exist at several addresses.

Finally, plan for the catalogue changing. Products get discontinued, categories get merged, seasonal ranges come and go. Decide a policy now for what happens to those pages, because the default behaviour on most platforms is to delete them, and deleting a page that had rankings and links throws away work you already paid for.

Fix the things that prevent indexing before you improve anything else, because a page a search engine cannot reach or will not index cannot be optimised into ranking. On most stores this is a short list of specific problems rather than a general audit.

Start with indexing coverage. Compare the number of pages you want indexed against the number actually indexed. Large gaps in either direction are the signal. Too few means something is blocking access. Too many usually means filter combinations, session parameters or duplicate paths are generating pages you never intended.

Then check the obvious blockers. Robots directives that exclude sections by accident, often left over from a staging site. Noindex tags applied broadly by a theme or an app. Pages that require a location or currency selection before they render. Content loaded only after a script runs, which some crawlers handle poorly.

Duplicate content is the next tier. The same product available at multiple URLs, with and without a trailing slash, with tracking parameters, or reachable through more than one category path. Canonical tags tell search engines which version is the real one, and getting them right resolves most of this. Check that the canonical on each page points where you think it does, because platform defaults and apps both interfere with them.

Speed matters, and it matters most on mobile product pages. The usual causes on a store are heavy images, too many third-party scripts from apps, and themes doing more than they need to. Fix the images first, because it is the largest gain for the least risk. Then audit the apps and remove the ones nobody uses, since each one adds weight to every page.

Structured data deserves attention because it changes how your listings appear. Product markup describing price, availability and reviews lets search engines show that information directly, and it gives them a second reading of your data that should agree with your feed.

Finally, make sure sitemaps are accurate and updated automatically, and that internal links do not point at redirects or dead pages. Both are cheap to fix and both waste crawl attention.

Assign every search in your map to a specific page type, because the most common reason a page fails is that it is the wrong kind of page for the query. Search engines are matching intent, and a product page cannot satisfy a search that wants options.

Four page types cover almost everything a store needs. Collection pages answer searches for a type of product, where the shopper wants to compare and choose. Product pages answer searches for a specific item, model or name. Guide pages answer searches where the shopper needs to understand something before choosing, such as how to pick a size, what the differences between materials mean, or what to look for. Comparison pages answer searches between two specific options a shopper is actually deciding between.

Collection pages carry more of a store's search traffic than any other type, and they are usually the least developed. A page that lists products and nothing else has nothing to rank with. Adding a genuine introduction that explains how to choose, the differences between the options, who each suits, and the questions buyers ask turns a listing into something worth ranking.

Product pages win specific searches and lose general ones. The pages that succeed are the ones with something unique on them: your own photography, detail that the manufacturer's description omits, real answers to buyer questions and genuine reviews.

Guide pages are where stores compete with publishers, and they only work if they are actually useful. A guide written to hold keywords is obvious to everyone including search engines. A guide written by people who sell the products every day and know what goes wrong is worth reading, and it earns links that catalogue pages never will.

Comparison pages should compare options a customer genuinely chooses between within your own range. Two of your products, two materials, two sizes, two approaches. That is useful and it is honest, and it converts because the reader is at the decision point.

One page per intent. Two pages targeting the same search compete with each other and neither wins.

Create content only where a catalogue page genuinely cannot answer the search, and make each piece something you would send a customer. Most store blogs fail because they are built to publish rather than to answer.

The useful pieces come from three sources. Questions your customers ask before buying, which you already know because your team answers them daily. Questions they ask after buying, about use, care, compatibility and problems. And decisions they have to make that your category pages cannot resolve, such as choosing between two approaches or working out what size or specification suits their situation.

Sizing, fit and compatibility content deserves special mention because it does two jobs at once. It ranks for a real search, and it reduces returns, which is worth money on its own. A store that publishes a genuinely careful sizing guide often finds it becomes one of the most valuable pages on the site.

Write from experience. Stores have information publishers do not: what actually gets returned and why, which product suits which customer, what goes wrong, what people wish they had known. That content cannot be researched from a desk, which is exactly why it earns attention.

Link every piece to the products it discusses, in the natural place, with descriptive text. A useful article that does not connect to anything you sell is a missed opportunity for both the reader and the store.

Keep the publishing rate low and the quality high. A handful of genuinely good pieces a year beats weekly filler, and filler carries a cost, since a site full of thin pages is judged as a whole.

Refresh rather than accumulate. Update the pieces that already earn traffic, add what has changed, and remove what has become wrong. Old content that stays accurate keeps working for years, and updating an existing page is usually a better use of an afternoon than writing a new one.

Finally, put a real byline on the reasoning where it helps. Content that clearly comes from people who handle these products daily reads differently from content that does not.

Point your internal links deliberately at the pages you want to rank, because internal linking is the one ranking factor a store controls completely. Most stores link almost entirely by accident, through navigation and automated related-product widgets, and the result is that authority pools on the home page and the newest blog posts rather than on the categories that sell.

Start with the navigation, since it links from every page. Whatever is in the main menu receives the most internal weight, so it should contain the categories you most want to rank rather than a full mirror of your department structure. Footers can carry a second tier without cluttering the header.

Then link across the catalogue in ways a template would not. From a guide to the collections it discusses. From a collection to the more specific collections beneath it and to the guide that explains how to choose. From a product to the other products a buyer genuinely considers, and to the sizing or compatibility page that answers the objection.

Use descriptive link text. Text that names what the destination is about tells search engines something. Links that say click here or view more tell them nothing, and they are also worse for shoppers.

Pay attention to the pages that receive nothing. Deep products, older collections and pages several clicks from anywhere are common in large catalogues, and they are the pages most likely to be crawled rarely and ranked poorly. A links audit that lists pages by how many internal links point at them usually reveals a large group with almost none.

Be careful with automated widgets. Related product modules and tag clouds generate enormous numbers of links with no editorial judgement, which dilutes the signal from the links you chose. Keep them, but do not rely on them to do this job.

Finally, redirect rather than delete whenever a page goes away, and point the redirect at the closest equivalent rather than at the home page. Redirecting everything to the home page discards the value you are trying to preserve.

Focus on the small number of link sources a store can genuinely obtain, because most link building advice was written for publishers and does not apply to a shop. A store's realistic sources are its suppliers, its customers, its category, its own data and its physical presence.

Supplier and brand relationships are the easiest and most overlooked. Manufacturers frequently maintain lists of stockists, and getting listed is often a matter of asking. The same applies to trade bodies, certification schemes and any organisation whose membership you hold.

Your own data is the strongest asset most stores never use. You know what sells, what gets returned, what people ask about, how demand moves through the year in your category. Published carefully and without exaggeration, that is genuinely new information, and it is the kind of thing writers cite because nobody else has it.

Customers and community produce links when there is a reason. Sponsorships, local involvement, events, partnerships with people who make things with your products. These are slower and they produce the kind of mention that is worth having.

Content that earns links is not product content. It is the tool, the guide, the reference table or the piece of research that other people want to point at. One genuinely useful resource beats a year of articles nobody links to.

Digital public relations works for stores when there is a real story, and it fails when the story is invented. A pitch built around a made-up survey is transparent and it damages more than it helps.

Set expectations honestly. Links accumulate slowly, most outreach is ignored, and the compounding effect takes quarters to show. What matters more than volume is relevance, since a link from within your category is worth many from outside it.

And do not buy them. Paid link schemes create a risk that sits on the whole domain, and the stores that suffer most are the ones that had built something real alongside them.

Write pages that can be quoted, because an increasing share of buying research now happens inside generated summaries and assistant answers rather than on a results page. The work overlaps heavily with ordinary SEO, but there are differences worth acting on.

The first is structure. Systems that assemble answers pull specific passages, so a page that states its answer plainly in a short paragraph near the top is easier to quote than one that builds toward a conclusion. Put the direct answer first and elaborate afterwards, in every section.

The second is specificity. Generated answers favour content that contains concrete, checkable facts: measurements, materials, compatibility, care instructions, what is included, delivery terms. Vague marketing language contributes nothing, and it gets skipped in favour of a source that says something definite.

The third is consistency across the web. Assistants assemble a picture of your brand from many sources, so your product names, specifications and prices should agree wherever they appear: your site, your feed, marketplace listings, supplier pages, review platforms. Contradictions make you a less reliable source.

The fourth is structured data, which gives machines an unambiguous reading of your product facts rather than requiring them to infer from prose.

The fifth is that reviews and third-party discussion matter more here than they do in ordinary ranking, because these systems draw on what other people say about you as well as what you say about yourself. A brand nobody discusses anywhere is hard to recommend.

Measurement is genuinely harder. Referral traffic from these sources is small and hard to isolate, and a summary that answers the question may produce no click at all. Practical signals include whether your brand appears when you ask common buying questions in the assistants your customers use, whether branded search rises, and whether the traffic you do get from these sources converts, which it often does at a higher rate because the visitor arrives pre-qualified.

Do not treat this as a separate discipline. It is the same content work with a stricter requirement to be clear, specific and consistent.

Judge SEO by the revenue it produces from non-brand searches, because everything else is a proxy that can move without the business changing. Rankings, impressions and traffic are all useful for diagnosis and none of them pays for anything.

Split brand from non-brand first. Searches for your own name rise when your advertising works and fall when it stops, so leaving them in the total makes SEO look like it is following your ad budget. Non-brand organic revenue is the number that tells you whether this channel is growing.

Then track a small set of measures consistently. Non-brand organic sessions and revenue. Revenue by page type, so you can see whether collections, products or guides are carrying it. The number of pages receiving any organic traffic at all, which is the clearest measure of catalogue coverage. And conversion rate of organic traffic, which tells you whether you are attracting the right searches.

Watch page-level movement for diagnosis. A page that moves from just outside the top results into them produces a step change in traffic, so knowing which pages sit in that zone tells you where a small improvement pays most.

Set the reporting rhythm to match the speed of the channel. Monthly is right. Weekly reporting on SEO produces reaction to noise, and daily rank checking produces anxiety and nothing else.

Expect the timeline to be measured in quarters. Technical fixes can show quickly. New pages take months to settle. Competitive category terms take longer than that. This is a reason to keep the work continuous rather than running it as a project that ends.

When deciding what to do next, rank candidate jobs by likely revenue against effort, and give preference to work that helps many pages at once. Fixing a template affects every page using it. Improving one product page affects one page. That single principle will direct most of your effort correctly.

Finally, write down what you changed and when. Search results move for reasons outside your control, and without a record you will attribute those movements to whatever you happened to do that week.