Googlebot Follows Ten Hops. That Is Not Permission to Use Them

Three findings share one mechanism and could not differ more in what they cost you. A redirect loop means the page is unreachable: not slow, not demoted, gone, and the browser says so in plain language as “too many redirects”. A redirect chain means the page arrives, eventually, after a detour. And a link pointing at a redirect means the redirect is doing its job correctly while your link is out of date.

Only the first one is an emergency. Sorting out which you have is most of the work here, because the fix for each is a different edit in a different place.

Why this matters for AI traffic

Google is specific about its tolerance: “Googlebot can follow up to 10 hops in a ‘chain’ of multiple redirects.” That sounds generous, and the same document immediately takes it back: “we advise redirecting to the final destination directly. If this is not possible, keep the number of redirects in the chain low, ideally no more than 3 and fewer than 5.”

The reason is not a ranking penalty. It is the two sentences after: “Chaining redirects adds latency for users, and not all user agents and browsers support long redirect chains.” That second clause is the one that matters now. Googlebot is the most patient client that will ever request your page. The crawlers and fetchers behind AI answers are a newer, more varied population, and they were not all built to Googlebot’s tolerances. A chain your rankings survive is a chain a less forgiving fetcher may simply abandon.

A loop needs no such nuance. Nothing resolves it: no crawler, no browser, no fetcher. The page is off the web until you break the cycle.

What each of these three actually is

• A redirect loop. A sends to B, B sends back to A. The URL is unreachable for everyone. Critical, always, no judgment required.
• A redirect chain. A sends to B sends to C. It resolves, and every visitor pays the latency of every hop. Google follows it up to ten deep, advises fewer than five, and prefers you point A straight at C.
• A link pointing at a redirect. Your page links to a URL that answers with a 3xx. The redirect is not broken; it is working as designed. Your link is stale, and every visitor takes an extra round trip for no reason. The fix here is not a redirect rule. It is editing the link to point at the final destination.

That last one is worth dwelling on because it is the most commonly misread. Nothing is wrong with your redirects. Someone moved a page, the redirect caught it, and the old URL is still sitting in your own markup. Adding another rule fixes nothing.

When is it worth fixing, and when should you leave it?

Fix it when:

• It is a loop. That is not a judgment call, that is a page nobody can load.
• The chain is on a route people actually take, or one you want cited. Every hop is latency paid by a real client, and not all clients wait.
• Your own internal links point at redirects. Editing your markup to name the final URL costs nothing and carries no risk.

Leave it alone when:

• It is a two-hop chain on a page nobody visits. Redirect rules are load-bearing plumbing, and rewriting them to save one hop on an unimportant URL is how loops get created.
• The chain crosses systems you do not control (a CDN, a marketing tool, an old vendor domain). Confirm you can actually edit the first hop before you promise yourself a fix.

The honest tradeoff: chains are a tax, not a wound. Compressing every chain on a large site is a migration project with real risk of breaking something that currently works, and the reward is milliseconds. I would fix loops the day I found them, fix chains on the routes that earn money, correct internal links whenever I am editing the page anyway, and leave the rest.

Deciding which redirects deserve surgery is the judgment. The rule itself is one line, and Get AI Traffic writes it with set_redirect, which refuses to point a URL at a target that is itself a redirect and refuses one that points back at itself, so it cannot build the chain or the loop it is there to remove. Note the limit: it writes redirect rules. A link in your page that points at a redirect is markup, not a rule, and no redirect tool will fix it.

Frequently asked questions

What does “too many redirects” mean? Your browser followed a redirect cycle until it gave up. Almost always A sends to B and B sends back to A, often from two rules that each work alone, like a force-HTTPS rule and a force-www rule that disagree. The page is unreachable until the cycle breaks.

How many redirects in a row is too many? Googlebot follows up to 10 hops. Google advises redirecting straight to the final destination, and if you cannot, keeping the chain to no more than 3 and fewer than 5. Treat 10 as the cliff edge rather than the budget.

Do redirect chains hurt my rankings? Not directly, and that is the wrong thing to watch. The documented costs are latency for users and the fact that not every client supports long chains. Googlebot is patient. The fetchers behind AI answers are newer and more varied, and a chain is a chance for one of them to give up.

Should I use a 301 or a 302? A permanent redirect (301) shows the new target in search results; a temporary one (302) keeps showing the source page. If the move is real, permanent is what tells search engines to move with you.

My audit says a link points to a redirect. Do I add a redirect? No. That finding means your own markup names an outdated URL that something already redirects correctly. Edit the link to point at the final destination. Adding a rule would just be a second detour.

Sources

• Google Search Central: Site moves with URL changes
• Google Search Central: Redirects and Google Search
• Google Search Central: HTTP status codes, network and DNS errors