Ecommerce search work is different from every other kind because the site generates its own pages faster than anyone can write them. Filters, sort orders, variants, pagination and search results multiply into thousands of near-identical URLs, and most of the technical work is deciding which of them should exist at all. On top of that sits the ordinary content problem: category pages that say nothing, product pages copied from the manufacturer, and no page at all for the way customers actually describe what they want. An agency that only talks about content has not seen the crawl. An agency that only talks about crawling has not read the category pages. This page describes what the work is, in the order it usually pays, and how to test a provider before committing.
The category page is the money page
On most stores the pages that can rank for commercially valuable queries are category and subcategory pages, not individual products, because the shopper searching for a type of thing wants a choice rather than one item. Yet category pages are usually the emptiest templates on the site: a heading, a grid and nothing else. The work is to give each one a genuine reason to exist, which means the page matches a real way customers narrow their choice, carries copy that helps someone decide rather than a keyword paragraph pushed below the fold, and links sensibly to the subcategories under it. Building a page for every possible filter combination is the opposite mistake and produces thousands of thin duplicates. The judgement call, on every store, is which intersections deserve an indexable page and which should stay as filters, and that judgement is most of the value an experienced provider adds.
Product data, structured properly
Product pages need the facts a buyer compares on, written for this store rather than pasted from the supplier, plus structured data so search engines can read the price, availability, review rating and shipping details directly. Google documents the product markup it supports, including the properties for price, availability and returns, and getting these right is what makes a listing show useful detail rather than a bare title. Variants deserve particular care: a size and colour matrix can either collapse into one strong page or explode into hundreds of weak ones depending on how the store is built. Out of stock lines need a policy decided in advance rather than a redirect improvised each time, because deleting a page that has earned links throws that value away. Ask a candidate agency how they would handle discontinued products on your specific platform and listen for a considered answer rather than a general one.
Speed, crawl budget and the boring half
Large stores routinely spend most of their crawl allowance on URLs nobody should ever see: filtered views, session parameters, internal search results and infinite pagination. Fixing that is unglamorous and often the highest-return work available, because it is the difference between new products being discovered in a day and in a month. Page speed matters for the same reason it matters everywhere on a store, which is that slow pages lose buyers regardless of ranking. The published Core Web Vitals metrics give a common language for that conversation: loading, interaction responsiveness and layout stability, measured on real visits rather than in a lab. An agency that can read those field measurements and tell you which template is responsible has done this before. One that sends a screenshot of a lab score has not.
Judging the provider
Ask for a store they worked on where you can see the category structure they built, then look at it as a shopper. Do the categories match how you would narrow a choice? Does the copy help? Are there obviously duplicate pages one click into the filters? Next, ask what they would do in the first month on your store specifically, and check whether the answer requires access to your crawl data, because a genuine answer does. Ask who writes the product copy and whether it is in scope or billed separately, since content volume is where ecommerce retainers quietly balloon. When a store buys this as a package it is usually sold as ecommerce SEO services with a fixed monthly figure, so pin down how many pages that figure covers and what happens when the catalogue grows. Finally, own the analytics and Search Console accounts yourself.
Questions people ask about seo for ecommerce websites
Should filtered pages be indexable?
Some of them, and the selection is the whole skill. A filter intersection deserves an indexable page when people genuinely search that way and the resulting page has enough distinct inventory and copy to be useful. Everything else should be crawlable for navigation but kept out of the index. Indexing every combination creates thin duplicates that compete with each other, and indexing none of them leaves obvious demand on the table.
What should happen to a discontinued product page?
Decide a policy rather than improvising. If a direct replacement exists, redirect to it. If the line is simply out of stock and returning, keep the page live and say so, since removing it discards any links and history it has earned. If nothing replaces it and it will never return, redirect to the parent category rather than the homepage, and only serve a not found status when there is genuinely nowhere sensible to send the visitor.
Is manufacturer product copy a problem?
It is a missed opportunity more than a penalty. Every competitor selling the same item has the identical text, so nothing on the page gives a search engine a reason to prefer yours. Rewriting the entire catalogue is rarely realistic, so prioritise: rewrite the products that already earn impressions and the ones with the best margin, and add genuine reviews, sizing detail and photography everywhere else.
How long before ecommerce search work pays back?
Technical fixes to crawling and indexing can show movement within weeks because they change what is discoverable rather than what is competitive. Category and content work behaves like any other content investment and should be planned across quarters rather than months. Sequence accordingly: take the technical wins first so the later content work is discovered quickly, and hold paid search steady while the organic layer builds.