Most online stores don’t lose traffic because their products are wrong. They lose it because their category pages read like filing cabinets, their filters spawn ten thousand near-identical URLs, and the only page with real writing on it is the About page. Ecommerce SEO is the work of fixing that — turning a catalogue into something search engines can crawl cheaply and shoppers can trust quickly.
This guide walks through how we approach ecommerce SEO for Canadian retailers: the architecture decisions that matter before any keyword is targeted, how to write category and product pages that actually convert, what to do with faceted navigation, and how to measure the work in revenue instead of rankings.
Why ecommerce SEO behaves differently
A service business has maybe forty pages. A store with 3,000 SKUs, five filter types and three sizes per product can generate several hundred thousand crawlable URLs without anybody deciding to make them. Google will not crawl all of them, and it certainly will not index all of them well. So the first job is not “write more content.” It’s deciding which pages deserve to exist.
The second difference is intent spread. Someone typing winter running jacket is browsing. Someone typing Saucony Triumph 22 mens 10.5 has a credit card open. Those two people need completely different pages, and stores that collapse both into one product grid lose the first shopper to a competitor’s guide and the second to Amazon.
Architecture: the part nobody wants to pay for
Site structure is the highest-leverage and least glamorous part of ecommerce SEO. Every product should be reachable from the homepage in three clicks or fewer. Every category should sit under one canonical parent, not float in a tag cloud. And the URL should read like a path a human could describe out loud.
A structure that holds up at scale
- Top-level categories map to how customers shop, not how your supplier ships. “Running” beats “Footwear – Athletic – Category 4B.”
- Subcategories target the modifier searches with real volume: trail running shoes, waterproof running jackets. These are the pages that earn links and rankings.
- Product pages sit one level down and carry the specific model and variant terms.
- Editorial hub lives beside the catalogue, not inside it, and links down into the categories it references.
Keep URLs shallow and stable: /running-shoes/trail/ rather than /collections/all/products/filter?type=trail. When a category is retired, 301 it to the closest surviving parent. Nothing bleeds authority faster than a store that quietly 404s two hundred seasonal collections every January.
Internal linking is your budget allocation
Crawl budget and link equity flow along internal links. If a subcategory only exists in a dropdown menu, it gets a fraction of what a subcategory linked from the homepage body, from related products, and from three blog posts receives. Decide which twenty pages carry your revenue and link to them deliberately from everywhere it makes sense.
Faceted navigation without the sinkhole
Filters are the single biggest technical risk in ecommerce. Colour, size, price, brand and availability combine multiplicatively, and each combination is a URL. Left unmanaged, Googlebot spends its crawl on size=9&colour=navy&sort=price-desc and never revisits the pages you care about.
| Facet type | Example | Treatment |
|---|---|---|
| Demand facet | Brand, material, use case | Indexable static URL with unique copy and title |
| Refinement facet | Size, colour on apparel | Canonical to parent, or block via robots rules |
| Session facet | Sort order, page view mode | Never linked as crawlable href |
| Pagination | Page 2, 3, 4 | Self-canonical, keep crawlable, keep products linked |
The rule of thumb: a facet earns an indexable page only when people search for that combination. Womens waterproof hiking boots deserves a page. Womens waterproof hiking boots size 7 in navy sorted by price does not. Pull your facet decisions from search volume data, not from what the platform happens to generate.
Category pages that rank and sell
The average category page is a grid, a title and nothing else. That page has no chance against a competitor whose page explains the difference between neutral and stability shoes, shows the three bestsellers with reasons, and answers the four questions the sales floor hears every day.
Two or three sentences that confirm the shopper is in the right place and mention the defining attribute they searched for.
Badges that do real work: in stock, ships from Ontario, free returns. These lift click-through and reduce bounce.
Buying guidance, sizing notes, comparison table, and FAQs. This is where the long-tail rankings come from.
Links to the two or three sibling categories a shopper might have meant instead. Fewer dead ends, more indexed depth.
Write the copy for the person, then check it covers the vocabulary the query data shows. Not the other way round. Stuffed category text reads exactly like what it is, and it doesn’t convert.
Product pages: the details that move rankings
Manufacturer descriptions are duplicated across every retailer who sells the item. If your product page carries the same 90 words as forty other stores, there is nothing for a search engine to prefer. The fix isn’t heroic — it’s specific.
- ✓ Rewrite the first paragraph in your own words, including the use case and who it suits.
- ✓ Add a short “what’s different about this one” section comparing it to the nearest alternative you stock.
- ✓ Publish real specifications in a table — weight, dimensions, materials, warranty, country of origin.
- ✓ Collect and display reviews. They add unique text, answer objections and feed review rich results.
- ✓ Include a Q&A block sourced from actual customer service emails.
Handle variants deliberately. If size and colour matter to search demand, give each variant its own indexable URL with a distinct title. If they don’t, keep one canonical product URL and switch variants client-side. Mixing both approaches across a catalogue is how stores end up with 8,000 indexed pages and 400 that get traffic.
Structured data worth implementing
Product markup with price, availability, brand, GTIN and aggregate rating is the baseline. Add BreadcrumbList so the SERP shows your hierarchy, Organization for entity clarity, and FAQPage where you genuinely publish questions. Keep markup honest — mismatched prices between schema and page are a common cause of merchant listing suppression.
What good performance looks like
Ecommerce is where Core Web Vitals stop being an abstraction. Product images, review widgets, chat bubbles, three analytics tags and a sticky promo bar all compete for the same main thread on a mid-range phone on a suburban LTE connection.
The practical wins are almost always the same four: serve images in modern formats at the size they render, preload the hero image and the primary font, defer every third-party script that isn’t payment or consent, and stop shipping the entire component library to a page that shows a grid.
Content that feeds the catalogue
Editorial content earns links and captures the research phase. But it only pays when it links down into commercial pages. A guide to choosing a road bike should link to the road bike category, the two bikes it recommends, and the accessories a first-time buyer forgets.
Three formats consistently work for Canadian retailers:
- Comparison pieces — “X vs Y”, written honestly, including who should buy the cheaper one.
- Seasonal and climate guides — the Canadian angle is a genuine advantage. Winter, salt, humidity and shipping timelines are things offshore competitors write generically about.
- Sizing and fit content — the highest-intent informational query in most apparel and equipment categories, and the one that reduces returns.
One well-linked buying guide that ranks for a research query tends to be worth more than twenty thin posts written to hit a publishing quota.
Off-site: links that survive an algorithm update
Product pages rarely earn links on their own. Categories and editorial do. Focus link work on the pages that can attract citations — original data from your own sales, tools like a size calculator or cost comparison, and supplier or manufacturer pages that list authorised retailers. Local press and industry associations still hand out clean links to Canadian merchants who ask with something worth covering.
Skip anything sold by volume. A store that adds 500 links a month from directories will spend the following year cleaning them up.
Measurement: revenue, not rankings
Rank tracking on a 3,000-SKU store is noise. What matters is the trend in non-brand organic revenue, and the ratio between the pages you invested in and the traffic they now hold.
| Metric | Why it matters | Where to check |
|---|---|---|
| Non-brand organic revenue | Removes the flattery of people searching your name | GA4 with a brand-term filter |
| Indexed vs valuable pages | Exposes facet bloat and thin-page problems | Search Console page indexing report |
| Category page click share | Shows whether titles and snippets are competitive | Search Console, page-level |
| Crawl requests by URL type | Reveals where crawl budget actually goes | Server logs or crawl stats |
| Product page conversion by traffic source | Separates a ranking problem from a merchandising problem | GA4 landing page report |
A realistic first 90 days
- Weeks 1–2: full crawl, log sample, index audit. Decide which URL patterns should exist and which should be consolidated or blocked.
- Weeks 3–5: implement facet rules, fix pagination and canonicals, clean redirect chains, ship performance fixes.
- Weeks 6–9: rewrite the top 20 category pages and the top 100 product pages by revenue potential. Add schema across templates.
- Weeks 10–13: publish the first editorial cluster, build internal links from it, start outreach on the two assets worth linking to.
Movement on long-tail product queries usually shows inside six to eight weeks. Competitive category terms take longer — three to six months in most Canadian verticals, more if the SERP is dominated by marketplaces.
Frequently asked questions
How long does ecommerce SEO take to pay back?
Technical fixes and product page rewrites often show within two months because the demand already exists. Category rankings against established competitors typically take three to six months, and the compounding effect keeps building after that.
Is Shopify good for SEO?
Yes, with caveats. It handles the basics cleanly but forces /collections/ and /products/ into URLs and generates duplicate product paths through collections. Both are manageable with canonical rules. Platform choice matters far less than architecture and content quality.
Should every product have unique copy?
Every product that could realistically earn search traffic, yes. For a 10,000-SKU catalogue, prioritise by revenue and search volume and work down the list; templated copy with genuinely unique specification data is an acceptable floor for the long tail.
Can a small store compete with Amazon?
Not on generic head terms. It can win on specific models, niche use cases, expertise-led guides and local intent. That’s where most of the profitable volume lives anyway.
What about AI search results?
The same fundamentals apply, plus clean structured data, clearly chunked answers and third-party corroboration. Engines that summarise cite sources that are explicit and consistent about specifications, price and availability.
Want to know which pages are holding your store back?
SEOelinks runs a free technical and content audit on your catalogue — index bloat, facet handling, page speed and the category pages worth rebuilding first. No lock-in, month to month, cancel anytime. Request your free audit.