302 to 301 redirect

[Tiggerito]

301 Redirect is more powerful than 302 redirect because of full transfer of backlink and Page rank .i do not think ranking would be suffered from doing 301 redirect .

301 does lose some Page Rank according to many, including Google. Just Google phrases like "google 301 redirect loss".

We have to remember that a 302 has a different purpose, and this is why it's use is generally bad for SEO. 302 is a Temporary Redirect which tells the reader... "I'm sending you to a new page, but is it only for a while. Normal service will be resumed in a while". Many use 302s when that is not the true message.

@deepsand highlighted a real reason for 302s (or triggered it in my mind). You may have to use a go between phase with your website upgrade when you do want to do Temporary Redirects. 302s are good then. I've rarely seen the case for this though.

[deepsand]

301 does lose some Page Rank according to many, including Google. Just Google phrases like "google 301 redirect loss".

According to Cutts, who, after admitting that he doesn't know all things Google, supposedly did some checking with others there, and reported that there was some dilution. However, he never explained why that should be the case, given that PR is a probability and the probability of a user who clicks on a link to a 301ed page has a 100% probability of requesting the new target page! From a mathematical standpoint, there should be no dilution.

We have to remember that a 302 has a different purpose, and this is why it's use is generally bad for SEO. 302 is a Temporary Redirect which tells the reader... "I'm sending you to a new page, but is it only for a while. Normal service will be resumed in a while".

Is that necessarily bad? Scraping my memory for any "official" Google mention of 302 handling, and coming up empty. I do, though, from personal experience know that pages which are redirected via a 302 are still indexed by Google.

@deepsand highlighted a real reason for 302s (or triggered it in my mind). You may have to use a go between phase with your website upgrade when you do want to do Temporary Redirects. 302s are good then.

Which was why I said that such were "eminently" logical under certain circumstances.

In the best of all worlds, one would always have ample time and resources for mapping all redirects in advance of transitions; not so in real life.

In the case where such is not well prepared for in advance, allowing requests to now non-existent pages be handled by a custom 404 routine, and have it return a 302, is a reasonable and acceptable interim solution.

[Tiggerito]

From a mathematical standpoint, there should be no dilution.

Good point.

I heard the PageRank algo has a built in juice leak between links. I think this is to stop it going into a +ve feedback loop type thing. In normal links on a page you will probably never get a 100% probability for a click, but as you say, a 301 is 100% guaranteed.

Maybe the leak is a form of punishment (wasting a Google request, their time and visitors time) or it could be simply just how the algo panned out!

Is that necessarily bad? Scraping my memory for any "official" Google mention of 302 handling, and coming up empty. I do, though, from personal experience know that pages which are redirected via a 302 are still indexed by Google.

You removed my last sentence from your quote, "Many use 302s when that is not the true message." which changing what I was saying. I'm not saying it is "necessarily bad", but it can be used wrongly.

I'd agree that it seems Google keeps the source of a 302 in its index, along with any ranking data. I'd assume that would also mean the destination does not get any link juice from it. You can't have both.

In the case where such is not well prepared for in advance, allowing requests to now non-existent pages be handled by a custom 404 routine, and have it return a 302, is a reasonable and acceptable interim solution.

I also agree here. A perfect use for 302s.

[deepsand]

I heard the PageRank algo has a built in juice leak between links. I think this is to stop it going into a +ve feedback loop type thing. In normal links on a page you will probably never get a 100% probability for a click, but as you say, a 301 is 100% guaranteed.

Given that the damping factor d is less that 1, a "positive feedback loop" is not possible.

As for "leakage," I've heard that used re. two different things, both of which are erroneous.

The first is the notion that whatever PR accrues to a target page is lost by the source page. What actually happens is that any increase in the PR of the target page ever so slightly reduces that of each and every other page in the calculation matrix, so that the grand sum of all remains equal to 1.

The other relates to the notion that pages that have no OBLs, so-called dangling nodes, cause PR to evaporate because the only sub-expression of the algorithm that exists for it is the first, i.e. (1-d)/N. This is, in actuality, immaterial.

Zero terms are of no consequence to the calculations if using simultaneous equations or matrix algebra. And, if using reiteration, any "missing" or "extra" PR is adjusted for at the end when all of the individual PR values are normalized so that their sum equal 1.

You removed my last sentence from your quote, "Many use 302s when that is not the true message." which changing what I was saying. I'm not saying it is "necessarily bad", but it can be used wrongly.

My apology for any such appearance, as that was certainly not the intent. I was just cutting to the chase re. a 302 not necessarily causing a loss, based on my experience that it is the target of such whose page is cached, and therefore presumably gets a benefit similar to or the equal of a 301.

I'd agree that it seems Google keeps the source of a 302 in its index, along with any ranking data. I'd assume that would also mean the destination does not get any link juice from it. You can't have both.

Yes, the page that is 302ed can still be found in Google's indices via the site: operator. Aside from that, though, other data returned via info:, such as link:, related: and the cache, relate to the target page. The only oddity is the link: returns a query string with an string of "nonsense" characters, such as -4jpZGrfl2YJ, inserted between it and the target URL, with the message "did not match any documents."

[Tiggerito]

I'd rep you for that, but I've been told to spread it around :-O

[deepsand]

No problem; it's the thought that counts.

I've run into that often enough myself for some time now, such that I suspect that the minimum required number of rep. posts between going to the same member has been silently increased.

As for the odd behavior noted re. links, I'm still trying to puzzle that out. When I've a moment, I'll look to GWMT for signs of links being passed through via 302ed URLs.

[Tiggerito]

Back to the OP. Matt Cutts just released a video where he recommends you don't combine 302s and 301s...

http://www.youtube.com/watch?v=PQCrRwTjkO8

[deepsand]

That's from 19 AUG 2011 - one that I'd previously seen - and, typical Cutts' obfuscation.

He doesn't really say that mixing is definitely a problem - which, BW, need not necessary be done via 302s - or that such concatenations are a problem, but only that Googlebot might give up with more than 5 concatenated re-directs.

It's a simple case of how long G'bot will wait for a response from the end of a chain.