Redirects, www and the canonical host

Most sites are reachable at four addresses: http://example.co.uk, http://www.example.co.uk, and the HTTPS versions of both. Search engines treat them as different pages unless three of them redirect to the fourth. The Web facet follows both the HTTP and HTTPS trails and shows every hop.

What good looks like

• Exactly one canonical host, usually https://www.example.co.uk or https://example.co.uk; pick one and keep it.
• Every other variant answers a single 301 straight to the canonical URL, keeping the path: http://example.co.uk/about → https://www.example.co.uk/about.
• The canonical host answers 200 with an HSTS header.

Common problems

• Chains: http:// → https:// → https://www. costs an extra round trip on every first visit and dilutes link signals. Redirect in one step.
• 302 instead of 301 for a permanent move: browsers and search engines cache it less and keep asking.
• Certificate only for one host: https://example.co.uk fails the TLS handshake before it can redirect. The certificate must cover both names (or use a wildcard).
• Loops: usually a CDN and an origin server each redirecting to what the other considers canonical. The trail stops when a URL repeats.
• Redirecting to a path without the request path: sends deep links to the home page.

Checking after a change

Run the Web facet with ?fresh=1 to bypass the ten-minute cache and read both trails. The HTTP trail should be one hop; the HTTPS trail should end at the canonical host with status 200.