https://example.com/about, https://www.example.com/about, and http://example.com/about can all serve the exact same content, but search engines treat them as separate URLs unless you explicitly tell them otherwise. Add a trailing slash or two, and the count keeps growing. Nobody sets out to create duplicate URLs on purpose - it just happens when DNS, hosting, and CMS defaults aren’t aligned from day one, and it usually goes unnoticed until someone compares two Search Console properties side by side.
Why the same page ends up with several URLs
Most hosting setups resolve both www and non-www by default, and many also serve both http and https without redirecting one to the other. CMS platforms add their own quirks - trailing slashes, case sensitivity, or index.html suffixes that technically load fine but count as separate addresses to a crawler. None of this is visible to a human clicking around the site; it only shows up when you check how search engines and analytics tools see the URLs.
Why duplicate hosts actually hurt
- Backlinks and social shares scatter across different hosts instead of consolidating authority on one URL.
- Crawl budget gets spent re-fetching near-identical content instead of new pages.
- Analytics fragments into what look like multiple “sites”, making traffic and conversion numbers harder to trust.
- In the worst case, Google picks a different canonical than the one you intended, and your preferred URL doesn’t rank.
Trailing slashes, parameters, and other quiet duplicates
www versus non-www and http versus https are the most visible splits, but the same problem shows up in smaller forms that are just as easy to miss. /products and /products/ can be treated as different URLs depending on server configuration. Tracking parameters like ?utm_source= or session identifiers create a technically unique URL for every visitor. Sort and filter options on category pages (?sort=price, ?color=blue) multiply a handful of real pages into hundreds of crawlable variants. None of these need a redirect necessarily, but each one needs a canonical tag pointing back at the clean, primary version.
Pick one version and commit to it
Start with https everywhere - there’s no good reason to keep http reachable in 2026. Then choose www or non-www. Neither is objectively better for SEO; what matters is picking one and being consistent everywhere it appears: in ads, business cards, social bios, and every internal link on the site. Document the decision somewhere your team and any future developer can find it.
Redirects and canonical tags do different jobs
At the server or CDN level, every alternate host and http request should 301-redirect to the single preferred URL - this fixes the problem before a crawler or visitor ever loads the wrong page. On top of that, every page should carry a rel=canonical tag pointing at that same preferred URL, in its own <head>. Canonicals are a hint that helps most when parameters or near-duplicate content exist (like ?sort= or ?ref= variants); redirects are the fix for the host-level problem itself. Use both - they cover different situations.
Cross-domain canonicals are a related edge case worth knowing about: if a press release or a guest post is genuinely published on two different domains on purpose, a canonical tag on the syndicated copy can point back at the original, telling search engines which one deserves credit. It only works if both sides cooperate - a site that republishes your content without adding that tag will still compete with the original for the same rankings.
A real example: two hosts quietly splitting authority
A site launches with both https://example.com and https://www.example.com resolving without a redirect between them. Marketing links to the www version in ads, the developer’s internal links point to the non-www version, and a few backlinks from press coverage use whichever one a journalist happened to type. Six months later, neither version ranks as well as it should, because links and signals that should reinforce one URL are split roughly in half across two. Fixing it later means auditing every internal link and hoping external sites update their links too - much cheaper to decide this before launch.
To check whether this is already happening to you, type all four combinations of your domain (with and without www, http and https) into a browser and watch what happens - if any of them loads without redirecting to your preferred version, you have the problem. Search Console will also show it indirectly: if you have properties set up for more than one host variant, compare their Performance reports for overlapping queries. A quick curl -I against each variant from the command line shows the exact redirect chain (or lack of one) without relying on a browser’s cache.
A short checklist to lock this down
- Decide the primary host (www or non-www) and enforce https across the board.
- Configure 301 redirects at DNS, CDN, or hosting level for every alternate variant.
- Add a self-referencing rel=canonical tag to every page, pointing at the preferred URL.
- Update the sitemap, internal links, and the Search Console property to match the chosen host.
- Check ads, business listings, and social profiles for stray links to the non-preferred version.
Killer Click bakes this decision into every launch checklist, so www, https, and canonicals get settled once during setup - not patched months later after ad spend and backlinks have already split across two versions of the same site, and not left for a developer to guess at from whatever the previous host happened to configure by default.