One product in five categories — duplicate problem?

'Q: My product appears under multiple categories. Is that creating duplicate content issues?'

Short answer: Only if each category generates a different URL for the same product. One canonical URL fixes it.

The longer version: showing a product in five category listings is normal merchandising and totally fine. The problem appears when your platform mints a unique product URL per category — /sale/widget, /new/widget, /gifts/widget — all the same product. Now you have five competing URLs splitting signals.

Fix: every product resolves to one canonical URL regardless of which category the user clicked from. The product page's canonical tag points to that single address; the rest either don't exist or canonical to it.

For breadcrumbs and structure, assign one primary category as the canonical path even though the product appears in several listings. That keeps your hierarchy clean and your relevance focused.

Rule of thumb: a product can live in many lists but must have exactly one address.

Got a duplicate-content question? Drop it.

Which filter combinations should I let Google index?

'Q: I want some facet pages indexed but not all. How do I decide which?'

Short answer: Index a facet only when its combination has real search demand and unique inventory.

The longer version: don't decide by gut — decide by demand. A facet page earns indexing if it passes three tests:

— People search for it. 'Nike running shoes' has volume; 'Nike running shoes size 10.5 blue under $73' does not.
— It has enough products to be a satisfying page (a one-product facet is thin).
— It's not a near-duplicate of an existing category.

Generally: single-select on high-demand attributes (brand, type, primary color) — index. Multi-select stacks, price ranges, sort orders, in-stock toggles — never index.

Practical setup: maintain an allowlist of indexable facet patterns. Everything else gets nofollowed internal links plus a robots disallow on the param, so crawlers never even reach the combinatorial explosion.

Rule of thumb: if you'd write a meta title for it without cringing, it can be indexed.

Got a faceted-nav question? Drop it.

Can I canonical my product to Amazon's listing?

'Q: I sell the same product on my site and Amazon. Should my page canonical to the Amazon URL to avoid duplication?'

Short answer: No — never canonical your own pages to a marketplace. You'd hand them your rankings.

The longer version: a cross-domain canonical tells Google 'the version on that domain is the real one, index that instead.' Point it at Amazon and you're voluntarily de-indexing your own product page. People do this by accident through SEO plugins or feed tools — check yours.

Selling identical manufacturer-described products is fine; you're not 'duplicating' in a harmful way. To stand out:

— Rewrite the manufacturer's boilerplate description (everyone else uses the same feed text).
— Add your own photos, Q&A, and reviews.
— Build internal links and earn external ones to your URL specifically.

That's how a smaller store outranks a marketplace listing for long-tail product searches.

Rule of thumb: your canonical should point to a URL you own and want to rank. Full stop.

Got a canonical question? Drop it.

Q: 'We noindexed faceted pages but traffic didn't move. Did we waste our time?'

Short answer: Not if you watched the right metric — crawl budget, not rankings.

The longer version: A mid-size outdoor gear store had ~480k indexable URLs, mostly color/size/brand filter combinations. They added noindex,follow to any URL with 2+ active filters and blocked the worst parameter combos in robots.txt. Rankings on money pages were flat for the first 6 weeks — which scared them. But Googlebot requests to category templates jumped from 38% to 71% of total crawl, and new products started getting indexed in 2 days instead of 11. Three months later, category page sessions were up 34% because Google was finally re-crawling the pages that mattered.

Rule of thumb: Faceted cleanup is a crawl-efficiency play first, a ranking play second. Measure log files, not just positions, for the first two months.

Got an e-com SEO question? Drop it.

—
Кстати, по этой теме сейчас открыты позиции: @careers_radar

Faceted nav: block crawling or just noindex?

'Q: I added noindex to my filter pages but Google still wastes crawl budget on them. What gives?'

Short answer: noindex stops indexing, not crawling — those are two different problems.

The longer version: a noindex tag only works after Googlebot fetches the page and reads the tag, so every filtered URL still gets crawled. On a store with 40 filters across 200 categories, that's millions of low-value URLs eating your budget before anything gets de-indexed.

The fix is layered. Decide which facet combos have search demand (color, brand) and let those be real indexable pages. For the infinite junk (price sliders, multi-select stacks), block the parameter pattern in robots.txt so Google never fetches it. Use rel=nofollow on internal facet links so you stop feeding crawlers the path in the first place.

Rule of thumb: noindex for pages you want seen-but-not-ranked, robots block for pages you don't want crawled at all.

Got an e-com SEO question? Drop it.

What do I do with out-of-stock product pages?

'Q: A product is sold out. Do I 404 it, redirect it, or leave it up?'

Short answer: Depends on whether it's coming back — and here are the three branches.

The longer version: split by intent.

Temporarily out of stock (restocking soon): keep the URL live, keep it indexed, and add availability schema with OutOfStock. Surface a 'notify me' form and link to similar in-stock items. You keep the rankings warm for when inventory returns.

Gone for the season but returns yearly: same — never throw away a URL that earns links and traffic every Q4.

Discontinued forever: 301 redirect to the closest equivalent product or the parent category. Only 404 if there's genuinely no relevant replacement, because a 301 to an irrelevant page is a soft-404 trap.

Rule of thumb: redirect to relevance, not to the homepage — homepage redirects get treated as soft 404s.

Got an out-of-stock edge case? Drop it.