301 Redirect Means "Some Loss of PageRank" - says Mr Cutts

[tedster]

For quite a while now, I've been cautious about using 301 redirects instead of fixing the core issue - whether it's getting legacy backlinks changed or fixing website infrastructure problems. This advice was based on several things - avoiding chains of redirects, for example, or introducing a potential trust issue because 301s have been a spam tool so often. But mostly, it seems to me that some PR is lost in a 301.

Eric Enge just published a new interview with Matt Cutts that confirms this idea.

Eric Enge: Let's say you move from one domain to another and you write yourself a nice little statement that basically instructs the search engine and, any user agent on how to remap from one domain to the other. In a scenario like this, is there some loss in PageRank that can take place simply because the user who originally implemented a link to the site didn't link to it on the new domain?

Matt Cutts: That's a good question, and I am not 100 percent sure about the answer. I can certainly see how there could be some loss of PageRank. I am not 100 percent sure whether the crawling and indexing team has implemented that sort of natural PageRank decay, so I will have to go and check on that specific case. (Note: in a follow on email, Matt confirmed that this is in fact the case. There is some loss of PR through a 301). [my emphasis]

[stonetemple.com...]

I wouldn't say that means never use a 301 - it is one of the useful tools in our toolkit. But it does mean don't throw 301 redirects around like confetti. Do get legacy backlinks changed when a domain changes. And it's better to fix a server infrastructure issue directly, whenever you can, instead of just doing a patch job.

[tedster]

The 301 redirect does transfer PageRank across domains. As we now see confirmed in this interview, that PageRank apparently carries over with some loss, but at a low-level. In fact, Matt Cutts was discussing his experiment of moving his blog to a new domain for a while, and he said he found no problems.

[pageoneresults]

Matt Cutts: re: link juice loss in the case of a domain change: “I can certainly see how could be some loss of PageRank. I am not 100 percent sure whether the crawling and indexing team has implemented that sort of natural PageRank™ decay”

My Comment: In a follow on email, Matt confirmed that this is in fact the case. There is some loss of PR through a 301.

They are specifically talking about domain name changes and not on site 301 > 200 scenarios. Just want to clarify that.

There is further discussion at...

29 Tidbits from my Interview of Matt Cutts
[stonetemple.com...]

301 Redirect Means "Some Loss of PageRank" - says Mr Cutts

Title might be just a little misleading but, you know us SEO types. :)

[graeme_p]

They are specifically talking about domain name changes and not on site 301 > 200 scenarios. Just want to clarify that.

On the other hand he did no specifically say that it only applies to domain name changes: it could well be that on site 301s lose some PR as well.

[g1smd]

Yes, Matt was only talking about certain scenarios, but there are many others to consider.

- Single redirect within sub-domain (e.g. for 'named index' to '/' canonicalisation, or for moved content page)
- Single redirect within domain (e.g. non-www to www, and other types of canonicalisation)
- Single redirect cross-domain (e.g. example.com to newsite.com site move, or for .co.uk and .com canonicalisation)
- Multiple redirects within sub-domain (e.g. content page moved multiple times, or for stripping referrer/campaign parameters and loading a tracking cookie)
- Multiple redirects within domain (e.g. inadvertently chaining the rule for non-www to www canonicalisation and a rule for some other process)
- Multiple redirects across domains (e.g. for site that has moved multiple times)
- Combinations of the above making even longer chains (e.g. a 'named index' to '/' canonicalisation redirect within the old domain, followed by a non-www to www canonicalisation rule on the old domain, followed by a redirect to the same page on new domain, etc).

Add redirects that load cookies, and redirects that strip referrer/campaign parameters ahead of all this, and the visitor might be redirected four or five times from when they click a link, to when the content is actually served. I would not expect PageRank to survive that chasm.

[steveb]

Why is it everytime he says gray, black and white FUD threads appear?

"that sort" does not mean "all sorts".

[doughayman]

What about affiliate links, wrapped in redirects ? I create CGI and ASP wrappers around my affiliate links, which, when clicked, store pertinent information in local databases about my visitors, and then query a local DB of mine to obtain the destination affiliate URL, and perform a final redirect to the affiliate site itself (this, as opposed to just hardcoding the affiliate URL's directly in my pages). I always "nofollow" these links ? Are you saying that this is a bad practice ?

[tedster]

Doug, there should be no issue for you - you're not trying to send PageRank to affiliate site, you're trying to send your buyers there.

[Whitey]

I always "nofollow" these links ?

It's a good practice that often save's tanking the affiliate partner with a duplicate content filter. Seems like you know this anyway.

[doughayman]

OK, guys, thanks for your feedback here.

[Whitey]

Another interview with Matt Cutts revealing issues on being conservative with redirect application :

Is It a Bad Idea to Chain Redirects (e.g. 301-->301-->301)?

"It is, yeah."

Matt was very clear that Google can and usually will deal with one or two redirects in a series, but three is pushing it and anything beyond that probably won't be followed. He also reiterated that 302s should only be used for temporary redirects...but you already knew that, right? [seomoz.org...]

[Sgt_Kickaxe]

I've just 301'd an entire site to remove a forced /index.php/ in every page url and there has been no loss of search ranking (yet, 1 week later). Google already has all non /index.php/ pages returning ahead of their /index.php/ versions (Bing and Yahoo haven't even started yet).

Overall the traffic is exactly the same this week when compared to last, perhaps a little higher (1-2%).

If there is a loss of rank through a 301 it's either limited to inter-domain moves or a very small loss.

[g1smd]

He also reiterated that 302s should only be used for temporary redirects.

John Mu also recently stated to use a 302 redirect when you need to redirect the root "/" URL to an internal URL: [webmasterworld.com...]

[g1smd]

I've just 301'd an entire site to remove a forced /index.php/ in every page URL.

It can take a few weeks or more for bad effects to show up, but in your case each old URL maps to only one new URL. This should be absolutely fine.

[Whitey]

I think we got ourselves into trouble on a site many years back with 301 "re direct chains" introduced through indecisions about where content should sit . e.g. A > B > C > A

The site dropped out of the rankings - indeed it was nowhere to be found even though pages were cached. After several years the site came back slowly , although it's not fully indexed.

From Matt's remark's he says that Google will at some point stop following the redirect. But is there an issue with " trust " as well ?

[aspdesigner]

@jdMorgan: Agree that it's better to do a re-direct, than to split your PageRank into two pieces.

"You can tell by the fact that Matt couldn't immediately answer Eric Enge's question that this is not a major issue"

@tedster: That's the impression I got as well. It was like it wasn't an intentional reduction, but more like when he asked a tech about it, they told him "yes, but because of how we implemented it..."

OK, I think I may have figured-out what they are doing here!

Consider this from a programmer's standpoint. If you're working on the code / database for calculating PageRank, how do handle a 301? You can't simply replace the old domain name with the new one - you need to keep the original URL around, so when you find new links pointing to that URL, you know where to pass the PageRank to!

The simplest/cleanest way, I would think, would be to simply consider the originating page as a node (page) for PageRank purposes, basically, storing the 301 like a normal page (original URL) with a single outbound link (destination).

If you do it that way, then the PageRank damping factor comes into play - that small amount of "juice" that gets subtracted from the total amount ANY given page has available to pass to its outbound links. This lost PageRank is small - in the original PageRank docs published by Larry Page, I believe the example given was 15%.

As the PageRank calculations are going to apply this damping factor automatically when calculating the outbound PageRank for each node, it would probably not be considered worth the extra time & trouble to write "special-case" code, just to exclude the damping factor from 301 pages.

So if that is the case, the loss would be no different then that naturally occurring on a regular page with an outbound link!

Matt gave us a clue that also appears to support that hypothesis, in the language that he used -

"...that sort of natural PageRank decay..."

Matt could have just as easily said that the loss is no different than that normally occurring when PageRank flows through any page, and we would instantly know exactly what he was talking about. But where is the fun & confusion in that? ;)

"On the other hand he did no specifically say that it only applies to domain name changes: it could well be that on site 301s lose some PR as well."

graeme_p, I would say that is most likely the case, for several reasons...

1) If this is just natural PageRank decay as a result of the way 301's are implemented (as I theorized above, and supported by Matt's wording), then there should be no difference where it is pointing to, the damping factor would be applied to the node in either case.

2) PageRank is calculated on a per-PAGE basis, not per site. PageRank flows through links from one page to another, regardless of location. Internal links pass PageRank, just like external links do. There is no reason to believe that Google would go to all the extra trouble to "special-case" 301's to calculate PageRank differently.

3) People seem to think there is something special about a domain 301. But technically, there is no difference between an individual page 301, and a site-wide domain 301. When you do a site-wide 301, the server actually has to generate individual page 301's for each page someone requests from the original domain. As far as Googlebot is concerned, there is no difference either - all Google sees are the individual page re-directs in either case.

4) PageRank calculations are one of the most computationally intensive tasks Google has to perform. It just does not make sense to slow that process down with extra "special case" coding, just to handle internal & external 301's differently.

[peterdaly]

Having paid fairly close attention to Matt Cutts' comments about 301 redirects since 2003 or so, here's my understanding, although I can't point to exact quotes...

A 301 redirect is like a page with one link on it. (unless it links to a second 301)

Like PageRank passing in normal pages, a certain amount is not passed on. In PageRank version 1, 85% of the PR value could be passed while 15% is "lost".

My understanding is a 301 is subject to this exact same PR loss situation. So in theory, if the dampening has not changed, a 301 is 85% as good as a direct link.