IP-based auto-redirects that quietly hide your locales from Google

IP-based auto-redirects that quietly hide your locales from Google

Hypothesis: aggressive geo-redirection is one of the most under-diagnosed causes of locales never getting indexed — because the symptom (missing pages) looks like a crawl problem, not a redirect problem.

The mechanism. Googlebot crawls predominantly from US IP addresses. If your server force-redirects every US IP to /en-us/, the crawler can reach only the US locale. Your /fr/, /de/, /ja/ pages exist, are in the sitemap, carry perfect hreflang — and never get fetched, because the bot is bounced to English before it arrives.

The result: hreflang annotations referencing pages Google has never successfully crawled, producing return-tag errors that no amount of tag-fixing resolves.

The fix, in order of preference:

— Don't hard-redirect by IP at all. Serve the requested URL and suggest the local version via a dismissible banner
— If you localize the homepage by signal, key off Accept-Language as a hint, not a forced redirect, and never block direct access to a specific locale URL
— Ensure every locale URL returns 200 to a US-based fetch

Verification method: fetch a non-US locale URL from a US IP (or use the URL Inspection live test) and confirm it returns the localized page, not a redirect.

Caveat: some redirection is defensible for UX. The line we'd draw is that discovery paths must stay open — a bot, or a user with a direct link, must always be able to load any locale without being forcibly rerouted.