Discussion starter: an ecommerce SEO question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.
- eCommerce SEO
- Indexing
Variants by size and colour: one URL with selectors, separate URLs canonicalised, or separate indexable URLs? Which won more long-tail traffic?
Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.
1 reply
-
adminAdmin
The short answer is that none of the three models wins long-tail traffic by default, and no controlled public test exists that I can point to showing one beating the others across catalogues. What the mechanism does show is that the decision should be made per attribute, not per site. Colour often carries its own search demand and its own visual content, so indexable colour URLs can earn queries a single parent page struggles with. Size rarely does, so size usually belongs on a selector or a parameter URL that canonicalises back.
The option most people misread is the middle one. Separate variant URLs canonicalised to a parent are a crawl, feed and structured data mechanism. They are not a long-tail play, because a URL that Google folds into its canonical does not rank in its own right. If long-tail capture is the goal, the real choice is between one indexed page per product and more than one indexed page per product.
The three models are really two decisions
Every variant set involves two separate questions that often get merged: does each variant have its own crawlable URL, and which of those URLs should be indexed? Google supports both main patterns. Its product variant structured data documentation describes “single-page” sites, where there must be “only one distinct canonical URL for the overall
ProductGroup“, and “multi-page” sites, where there is no single canonical URL for the group. Either way, it says the site must be able to preselect each variant “directly with a distinct URL”.Model What Google can index Long-tail potential Main cost Usually fits when One URL, selectors only (JS state or #fragment)One page Only what the parent page’s text covers No distinct variant URLs, so weaker fit with variant markup and Shopping landing pages Almost never the best choice today Variant URLs (parameter or path) canonicalised to the parent One page, with variants crawlable Same as above, but signals consolidate cleanly Colour-specific queries depend on one page covering every colour Variants look alike, or demand sits at product level Separate indexable variant URLs Many pages Highest ceiling, if demand and distinct content exist Near-duplicates, crawl load, split signals, cannibalisation, thin pages when stock runs out Attribute has real search demand and visibly different content Fragments are the one clearly weak implementation. Google’s ecommerce URL structure guidance says URLs like
#blackand#whiteare treated as the same page because Google does not use fragment identifiers in indexing. The same page supports either a path segment or a query parameter per variant (for example/t-shirt/greenor/t-shirt?colour=green), and for optional variant parameters it recommends using the URL with the parameter omitted as the canonical.Why “canonicalised variant URLs” do not collect long-tail traffic
A canonical tag is closer to a filing instruction than a locked door. You label the parent as the master copy, Google evaluates that label alongside other signals, and if it agrees, the variant URLs become alternates that sit behind the parent. In practice that means Google clusters the duplicates, picks one representative, and ranks that representative. The variant URLs still matter for crawling, for Merchant Center landing pages and for
ProductGroupmarkup, but they are not competing for queries.So when someone asks “which model won more long-tail?”, the fair comparison is: does the parent page rank as well for “sage green linen shirt” as a dedicated sage green page would? That depends on what the parent actually shows. A parent page with the colour names in visible text, colour-specific images with descriptive alt text, and
variesByin its markup can rank for plenty of colour queries. A parent that only exposes the default colour in its HTML, with other colours loaded by script on click, gives Google very little to match against.Important distinction:
Crawlable variant URLs and indexable variant URLs are different things. You can have the first without the second, and on most catalogues you should have the first regardless of what you decide about the second.
The attribute test: does it change what people type and what they see?
The most useful rule I know is to split attributes by two questions. Does the attribute change the query (people search for it by name)? Does it change the page (images, product name, materials, price or availability)? An attribute that does both is a candidate for its own indexable URL. An attribute that does neither belongs on a selector.
- Colour in fashion, furniture, paint, phone cases: often changes both. People search “black chelsea boots” and expect to see black boots in the image.
- Size in apparel and footwear: usually changes neither. The page looks the same, and “size 10” queries are rare and fragmented. Size also drives most stockouts, which would create thin or empty indexable pages.
- Size in mattresses, bed linen, tyres, rugs: the exception. “King size quilt cover” or a tyre size is the query itself, and price usually changes with it.
- Material or finish: depends on the vertical. “Oak” versus “walnut” dining table can carry separate demand and different imagery.
This is why the hybrid most large fashion catalogues end up with is indexable colour pages, with size as a parameter canonicalised to the colour URL. It is not a Google rule. It is where demand and distinct content usually sit.
Hypothetical example:
A retailer has 2,000 parent products, an average of 5 colours and 7 sizes each. Indexing every size and colour produces 70,000 near-identical URLs, most of which differ only by a dropdown value. Indexing colours only produces about 10,000 URLs, each with distinct images and a colour name in the title. Keeping only parents produces 2,000. On a site whose crawl logs already show Googlebot rarely revisiting deep product pages, the jump from 2,000 to 10,000 might be justified by colour demand, while the jump to 70,000 almost certainly would not be.
What goes wrong with separate indexable URLs
The higher ceiling comes with specific failure modes. The first is that Google may not accept your canonical choices. If colour pages share the same description, the same review block and near-identical images, you will often see them under “Duplicate, Google chose different canonical than user” in the Page indexing report, which means Google has made the consolidation decision for you, often in favour of the page with the stronger signals (typically the parent, which holds most internal links).
The second is signal splitting. Reviews, links and engagement that once concentrated on one URL now spread across several. If reviews show only on one colour page, the others look thin to users and weaker in search. Showing the full product family’s reviews on every colour page is common, but then those pages converge again on shared content.
The third is commercial. Separate colour pages can make it harder for shoppers to compare colours, and they multiply merchandising work. A gain in long-tail sessions that comes with a lower add-to-basket rate is not a clear win. Measure revenue per product family, not sessions per URL.
What I would inspect before choosing
- Search Console queries with attribute modifiers. Filter queries by a regex of your colour names (and sizes where relevant) and look at which pages they land on and at what position. Impressions for colour terms on the parent page with poor positions suggest unmet demand. Almost no colour impressions suggests there is little to win.
- Keyword data at family level. Compare demand for the product term with demand for “colour + product” across your top families. Demand is usually concentrated in a few colours and a few categories.
- Rendered HTML of the parent page. Check whether every colour name, image and price appears in the HTML Google renders, or only the default variant.
- Crawl logs. Check how often Googlebot reaches product pages now. If it struggles with 2,000 parents, adding 10,000 colour pages will not be crawled well.
- Inventory behaviour. Check how often individual variants go out of stock. Frequent variant-level stockouts favour fewer indexable URLs.
- Merchant Center set-up. Variants are grouped with item group ID, and each variant should land on a URL that preselects it. Every variant
Offerin your markup still needs a price for merchant listing eligibility.
A decision framework
Condition Lean towards one canonical product page Lean towards indexable colour (or other attribute) URLs Attribute search demand Little or no modified-query demand Clear demand for “attribute + product” queries Visual and content difference Variants look nearly identical Distinct images, names, materials or prices Catalogue scale and crawl Large catalogue, weak crawl coverage Manageable URL count, healthy crawl frequency Stock pattern Frequent variant-level stockouts Variants stocked consistently Platform Default variant parameters that canonicalise to the product (common on hosted platforms) Ability to create separate products per colour and link them as a family Commercial risk Shoppers compare variants side by side Shoppers arrive with a variant already in mind Platform deserves a note. On Shopify, for example, the default
?variant=URLs canonicalise to the product URL, so indexable colour pages usually mean creating one product per colour and grouping them in the theme. That is a merchandising and inventory decision, not only an SEO setting, and it is expensive to reverse.How to test it rather than argue about it
Because the answer varies by category, test within your own catalogue.
- Hypothesis: indexable colour pages increase non-brand clicks and revenue for product families with measurable colour demand.
- Groups: pick matched sets of families (similar category, traffic, price and seasonality). Split colour pages out for the test group only.
- Primary metric: clicks and revenue for the whole family (parent plus colour URLs), not per URL, so you are not counting cannibalised traffic as a gain.
- Secondary metrics: impressions for colour-modified queries, Google-selected canonical in URL Inspection, crawl frequency on the new URLs, add-to-basket rate.
- Duration: at least 8 to 12 weeks, avoiding launches next to a core update, major promotions or seasonal peaks.
The result that would change the decision: if Google selects the parent as canonical for most new colour pages, or family-level clicks do not rise while conversion falls, revert to one indexed page and put the effort into making that page cover every colour properly.
My practical position
Give every variant a crawlable, preselecting URL, because Google’s variant markup and Shopping need it. Keep one indexed URL per product by default. Promote an attribute to its own indexable URLs only where Search Console and keyword data show demand for it and the page genuinely looks different, which in most catalogues means colour in visually driven categories and size only where size is the query. Then prove it on a subset before you roll it across the catalogue.
Need help with this on your own site?
Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.
Add a reply Cancel reply
Replies are open to approved members. Log in or apply to join.