Redirects and Loops

SEO Issues

[simka]

I'm interested to know of any experiences anyone has had with 'redirects and loops', especially regarding the 302. I understand a chain of redirects can be an issue but is this really uncommon? My main question evolves around a 'SEO' related flag that has appeared (in beta) via SEMRush. I am not doubting the metric but would like to know more about this issue - and if anyone else has come across it.

[Peter_S]

I don't understand what's it's about, but a "loop" can't be good, because it means it goes nowhere.

[lucy24]

By "loop" do you mean redirects that eventually end up back where you started? I hope not, since that will result only in an error message from the browser, and the user never gets to your site at all. (If it's an internal loop, the user eventually gets a 500-class error instead. But those are rare.)

A 302 is by definition a temporary redirect--i.e. it might be gone tomorrow--so it's not really worth stressing over. That's assuming it really is meant to be temporary, and isn't the result of a coding error on your end. A temporary redirect is often the default, so make sure you're not sending out a 302 when you mean 301.

For permanent redirects, think of the humans. Each 30x response means the browser has to send in a fresh request, which has to pass through an unknown number of transfer points, taking time and using resources, while the human users are twiddling their thumbs waiting for something--anything!--to happen. If your users are on satellite internet, they may literally be twiddling their thumbs, because there will be measurable time lapses. And if you're sending a 302 when it ought to be a 301, anyone who comes back five minutes later will get redirected all over again instead of making the right request the first time.

Meanwhile, from your end, why should your server have to process two or more requests, when a single one will do? Sure, the total number may not be vast. With human visits, redirects will generally affect only page requests; supporting files will be correct the first time, unless you've been making massive rearrangements to your image directories. Or--the other big “unless”--you've got systematic coding errors. The obvious one is image links hard-coded to give full absolute URLs, so if you've changed protocol, every single item gets separately redirected. Now you're talking about making your server do twice as much work as it needs to.

Uhm. What was the question again?