301 Redirect Test: How Much Link Juice are YOU Losing?
In a recent interview conducted by Eric Enge, Matt Cutts indicated that while 301 redirects are still the preferred method for migrating between pages and sites, a certain percentage of PageRank is lost through the redirect. This disclosure has stirred the pot in the SEO community whereas before it was assumed that a 301 would pass all of the link juice, now many are asking how much is lost? and what else can we do?.
Given that no definitive answer has been provided about the exact percentage lost, we decided that the only way to determine the true value of a 301 was to analyze a site that had recently changed domains. Fortunately, one of our clients moved domains in the beginning of 2010 and had implemented a 301 on their old site. We managed this account before and after the switch, so data from both the old and the new sites was available for analysis. The following case quickly outlines what weve seen with these domains in particular, and how the 301 is functioning.
In the first graphic we can see the Open Site Explorer data for both sites. This was where we started our research because we wanted to see how an independent tool was evaluating the two domains and what type of link juice was being passed through.
While it is clear that the 301 appeared to be passing significant value through this analysis, the most important factor was how the engines were treating the redirect.
This next graphic shows the total amount of organic search traffic to the two sites before and after the switch was made. As you can see, total traffic normalized fairly quickly once the switch was made.
It is important to note two things about the above graphic. First, the same number of keywords were tracked for each site keeping the KWs as a controlled variable. Second, the site being analyzed experienced a fairly significant jump in traffic around the holidays. So while this data is not ideal given the peaks in the graph, it does help demonstrate the overall results since the switch and the relative speed with which traffic stabilized.
The next graphs show the organic search traffic in Google, Yahoo!, and Bing for this sites top keyword. For reference, this word gets approximately 550k searches on exact match per month.
Clearly Google is passing value from the 301, and traffic numbers have actually improved since the switch. Yahoo! showed an initial spike, however traffic dropped significantly and has not returned to previous levels. Bing does not appear to be passing any significant value and traffic levels remain low.
So what does this all mean? With this quick analysis we were able to reach the following conclusions:
- It is safe to say that 301 redirects do pass value Our OSE analysis shows that the new domain has picked up a significant portion of the linking domains, as well as relative authority.
- In addition, we saw through our KW traffic analysis that the 301 seems to be working quite well in Google. We have not seen any substantial drop since the domain switch and in fact, traffic has steadily increased. Whether or not there has been a measurable drop, or PageRank decay, is hard to say at this point. Our data indicates that the 301 is passing value, and our traffic numbers say the same thing if there is some value being lost, it is not affecting the bottom line.
- The same cannot be said for Yahoo! and Bing. Our data showed that Yahoo! and Bing are not passing adequate value from the redirect and that total traffic numbers may be impacted by this drop. Although Bing has said that it can take two or three re-crawls for value to register, it does not look like the full value has been passed with the redirect. When analyzing the impact of 301s in your future campaigns keep this test in mind and be sure to check how each engines traffic is affected.
If you have any questions about the information provided above, please leave them below in the comments section. In addition, a great review of the entire interview with Matt Cutts (complete with helpful illustrations) can be found on the SEOMoz Blog.
Posted: 04.09.10
goodnewscowboy:
Really nice data Mark. Thanks for sharing this info.
I’m amazed at the Yahoo and Bing results. Who’d a thunk it could be so different.
Dena:
Oh my lawd Mark… These graphics were the exact things i’ve been expecting for a long while! There are so many myths about 301′s juice passing that i sometimes am frightened to 301 something at all. Thanks so much for the research :)
Robert:
My buisness partner wants to 301 redirect our blank.ca to http://www.blank.ca but it makes the site a mess for Safari users. The page doesnt recognize the css file anymore.
We are in Vancouver where Mac is very popular. Market share for Safari is just under %10 worldwide its gotta be twice as high here.
Is a 301 redirect to itself going to increase so much traffic that its worth totally ignoring mac users to the site? I say Hell No. If it were, why dont we see other companies forfeiting their mac customers for some seo points. Somebody help me.
Mark Lavoritano:
@Brian, @Dena & @goodnews – Thanks for checking the blog and for the feedback – much appreciated. I’m glad that the information was helpful.
@Robert – Interesting question. It sounds like there could be a coding concern there. If you have any questions about passing value through the 301, feel free to reach out!
Darren Shaw:
This is some useful research Mark. It’s good to see that Google does a good job of passing the juice, and it’s surprising how poorly Yahoo and Bing handle this.
I have heard that, at least with Google, it can depend on the content. If you buy a content and PR rich domain, then 301 redirect it to a site with unrelated content, the PR will not be passed.
Jason:
Great article Mark! The graphics and final synopsis were really helpful.
Rob Woods:
In response to Robert’s comment above, you definitely have a coding issue there. We redirect our example.com URI to http://www.example.com/ and have no issues with Safari rendering. I’d say fix your CSS and redirect the URI. Right now if you have incoming links to both versions you are splitting up your link equity. I would recommend always redirecting the various versions of your home page that people might link to, to the canonical version of the URL. Coincidentally, we’re in Vancouver as well…
Mark Lavoritano:
@Jason – Thanks for reading…really appreciate the feedback.
@Darren – Good point about purchased domains. That seems like the logical next-step for a test like this. It would be very interesting to see how much value was passed through 301 of that sort…which has me thinking…
Jeroen van Eck:
The question is did you just prove that PageRank hardly gets lost when 301 redirecting for Google? Or did you just prove that PageRank plays a negligible role in ranking for keywords when your keyword relevancy is high enough?
terry Van Horne:
Marc… have a few questions
were the redirects at the server level or page?
were wildcards used to redirect?
I have a few thoughts on this and believe based on what some others have said even Google may vary how much juice or text is passed from site to site. Hence the varying results reported by peeps generally thought to be quite knowledgeable. The reason for the questions is to ascertain how trustworthy Google may see based on the way the redirects are implemented.
Wendy Piersall:
I have seen different results based on variables not mentioned in this post.
When I moved an old site to a new domain (brand new), it took a full six months to recoup old traffic levels. So if your client moved at the beginning of 2010, there is still a chance that your traffic will nomalize in a few more months.
Additionally, I have moved content from a strong existing domain to an established domain. These redirects showed no down time at all in search traffic, and in fact Yahoo and Bing are sending more traffic to the new content vs. the old. I attribute this to two things: the fact that the new domains were already up and running and indexed, and the nature of the specific niche of the sites.
I have seen first-hand (and Matt Cutts has copped to this in the past), that search engines treat keywords in different niches on separate algorithims. So ‘work at home’ and ‘make money online’ are much more difficult to rank for than something like ‘coloring pages’ or ‘teacher worksheets’. I’m not sure if Yahoo and Bing have similar practices in place, but if the main keyword in your test gets over half a million searches, it’s safe to say it’s a pretty competitive one and may be influenced by this.
Andy FirstFound:
Wow. Bing really doesn’t bother with 301′s, does it?
Seriously though, great research. Thanks for putting this together, and have a Tweet on us!
Scott Skurnick:
Very interesting. We moved a very large site towards the end of last year and noticed the exact same thing with Yahoo. This helps verify my assumptions.
Four Pillars:
Thanks for this article – I’m switching my domain in mid-May so hopefully I will have similar results.
Rob W:
Scott
I agree that yahoo and msn likely pass less value through a 301 but if you only moved the site around 4 months ago you may need to wait longer to see the true effects. I’m seeing those 2 engines re-indexing very slowly right now. We took a site split into numerous subdomains back to just the www subdomains last fall and while goog re-indexed fairly quickly the other two haven’t finished recognizing the 301 redirects and flushing the old URIs out of their index even 6 months later.
richardbaxterseo:
I love SEO blog posts that share data and give real insight into the inner workings (and differences) between search engines. Enjoyable read, thank you!
pk_synths:
Nice data.
Was a domain move request submitted to Google Webmaster Tools before the domain was switched over? If so than that this could be affecting your 301 results as the domain switch request is a backend request. Could also explain why Yahoo and Bing didnt pick up on the new domain as I am not familiar with similar domain switch requests by those 2 engines.
Mark Simon:
This is a very good post however I believe this data shows more how traffic levels are affected with switching domains and 301 redirects. It is not really going into link juice lost/passed.
I personally would like to see a graph of your domain link metrics graphed similar to how you did the traffic data. Granted this data doesn’t change as often and you may not have it.
Thanks for sharing
Wiep:
Nice test, Mark!
Was the new domain a brand new one, or was it an established domain/ website too?
matt:
I’ve seen the same thing in rank data after a single page gets moved. We move some single pages to incorporate better keywords in the URL (and, more importantly, to get rid of some misleading keywords that confuse search-engine users). The old URL is 301′d to new, and all onsite links are changed to point directly to the new page. Within 2 weeks, usually faster, Google SERPs reflect the URL change. SERP rank is consistent. But Yahoo and Bing fare less well; typically, both these engines will drop the page entirely from their SERPs for up to 2 weeks. When the new URL eventually gets indexed, it ranks 5-10 places below the original. Of course rank is volatile and personalized, but this correlates strongly to the traffic data presented above.
James Green:
Love that you are looking at this in a data driven way rather than just anecdotal and especially thanks for sharing.
Ben Cook:
Sorry, but how are you deriving any useful conclusions from this data at all?
You don’t mention anything about whether or not you created new links, got new organic links, etc. The title is talking about how much Link Juice is passed but all but one of your graphs & data deal with traffic. As you mentioned in the post with seasonality, other things play into traffic which means that this really doesn’t prove much of anything.
There’s no where near enough control going on to isolate how much link equity is being passed. This simply shows that Google seems to transfer traffic fairly well over a three month period.
Ben Rush:
Good article, but I don’t think it actually answers the question it sets out to answer particularly well. All this proves is that juice is passed (which we knew already) followed up with a fairly broad statement that it seems to be quite a lot. What is quite a lot? 50%, 60%, 70%? There is to many variables that could be factoring into this research outside of just the 301 redirect.
The piece about Bing / Yahoo however is very interesting however and would be useful to revisit again to see if this has changed (i.e., they have traditionally been slower to pick up changes / new sites, pages etc).
seo agentur:
….I have seen first-hand (and Matt Cutts has copped to this in the past), that search engines treat keywords in different niches on separate algorithims……
that´s exactly what is was thinking first when reading this article! no doubt good analysis, but might vary from niche to niche. in order to have reliable results, tests should cover at least a couple of different niches and branches and therein also multiple sites, just to make sure…
wil reynolds:
Mark, you have big shoes to fill for blog post #2!!
@skizzo, I agree that the title might have actually been off a bit, because it really is more of a shoot out between how google, yahoo, and bing treat 301′s instead of the link juice lost.
g13 media:
Nice article Mark surprised MSN isn’t accepting the 301, wow. This really sucks because one will lose traffic if you change domain names.
Mark:
@terry Van Horne – Yes the redirects were done on the server level and I believe wildcards were used. Thanks for checking in!
@Wendy – Good point about the keyword niches. You are right that the search volume for KW may have impacted this test.
@richardbaxterseo – Thanks for the kind words – glad you liked the post!
@pk_synths – Yes the domain move was submitted to GWT. I agree that this could have affected our results. However, the fact that Yahoo and Bing do not have a similar process does not explain why they are not picking the site up, but the fact that we can point Google in the right direction may help explain why Google picked it up so fast.
@James Green – Thanks for checking the blog, appreciate the feedback.
@Ben Cook – Thanks for your feedback. I agree that this test was not a ‘perfect’ one, and the variables (ie seasonality) definitely could have impacted our results – However, we did make a point to highlight what value we were seeing from an independent tool, and how that was translating to the engines – and specifically how it impacted our traffic results. Although data like this is not all that common, I am hopeful that we may be able to analyze something similar with more controls. (and we’ll definitely be keeping an eye on this site to see if/when things change).
@Ben Rush – Thank you for the feedback. I’m not sure we’ll ever know the true percentage lost, but from this test we can see that Google looks like it is leaking less value than either Yahoo or Bing – which is definitely valuable data. I also agree about the variables in the test, however because we were working with real data, and not a test set, we were somewhat tied to the numbers and the lack of control – definitely something we will consider the next time around though!
Kenichi Suzuki:
Great analysis.
Would you let me introduce the article to my blog readers with translation? I’m a Japanese webmaster. Most of them are not able to understand your story in English(I am).
Branko:
Your Title is highly misleading. You cannot measure how much link juice you are passing, losing or gaining, unless you work at Google. As Ben pointed out, the only valuable takeaway from the results you are showing is that it takes longer for Yahoo and Bing to 301 your site to a new URL.
The results from the independent tool (SEOMoz) are a product of an algorithm that is based on certain assumptions. those assumptions could be better than your or my assumptions as they are based on a much larger set of data than we can put our hands on, but they are still assumptions that may have nothing to do with the way link juice/authority/relevancy are passed throughout the web.
It is nice though to see data/observation-based posts. They do stand out from the background noise of opinions that everyone has and some deem important enough to put on a blog post.
Directory Sieve:
nice analysis, i have done 301 redirects myself on my niche sites but never analyzed so deeply. however, i have never seen a rise in google organic traffic after the re-direction, actually its always a near about 20% decrease. Yahoo though drops a lot on search traffic but i have seen it come up after a long time. Unfortunately the sites that i did the re-direct did not received much traffic from bing so i don’t have any bing stats.
From your graphs, it looks like bing dose not values or even registers a 301 re-direct which is a pretty bad thing to do, surely many will lose faith in the bing search engine if Microsoft fails to add the value of 301 in bing algorithm :(
Nancy Hutchins:
I’m glad I found your article. I’m not a SEO pro by any stretch, but recently did a 301 re-direct from my old site and have been wondering about the effects. This is reassuring news. Thanks.
Shiraz:
hmmm, well… it’s nice… I was really curious about the link juice and the backlinks passing on to the new redirected domain…. thanksssssssss :)
Azzam:
I was truly concerned on this subject since I have a PR5 site with 200,000+ backlinks that I want to do a 301 redirect for and just do not want to lose any of the backlinks or PR5.
thanks, will keep on researching.
need more traffic:
That’s some good investigation!
My question is, what is you were to change the 301 of the old site to a another new site?
Would the new-new site do just as well and the new site above? Or would it not benefit since Google looks at 301 as a permenant thing and won’t re-follow the links and attribute them to the new 301?
sean fyresite:
Nice post, I have been trying to track down good info on 300 ridirects vs link juice for a while. thanks!
Todd:
I have actually just implemented a 301 redirect tonight, let’s see what happens.
Paul:
Great topic…I am having a HUGE problem right now with this…any ideas?
So I received a cease and desist order for my subdomain (sub1.mydomain.com).
I did a 301 redirect to the new sub domain (sub2.mydomain.com).
The rankings stayed the same and google picked up on this and everything was fine.
THE PROBLEM:
About a month later the old sub domain came back in the rankings and I had duplicate rankings which included the old sub domain and the new one. The new one got bumped down big time but the old one stayed in the top positions.
MY QUESTION
What do I do? I don’t want to lose my rankings but I can’t keep the original subdomain up because I will get sued. Any ideas why this happened? I did not remove or change ANYTHING!
leon:
thanks for the nice data
now i can safely use 301
Review Pedia:
Thanks for the article, one of my website is has little traffic from Google, but more traffic from Yahoo and Bing, so I wonder if I should do the 301 Redirect.
davidwayne:
Great article, I’ve been reading up on 301′s (since I’m debating doing it with my old site to my new domain name) and it seems like most sites have just walked in circles without giving a definitive answer. After reading this one, i’m going to go ahead and take the plunge and hopefully see a traffic spike on the new page.
Thanks again
Albanos Ismaili:
I can second that, but anyway great article.
I added on my Blogger Blog an old .de domain, and in a few days link juice was passed, and in the Google rankings the new site appeared.
Traffic is now not more, and not less.
Chris:
We recently switched domains and a different URL structure and we have experienced a 30% drop in our search engine traffic across the board. Based upon an SEO analysis we just did it has become clear that Google (and Bing) have not passed along all of our previous “SEO Juice” to our new website even though we have 301s set up for each of our old pages to our new pages. For the first 3 weeks of our new site we saw a steady drop in traffic until we hit the “bottom.” We have not yet started to go back up. We also dropped in Page Rank from a 4 to 3 and have not gone back up.
ebinx:
I would not trust SEOMoz tools entirely. The way they calculate Trust and authority is based on their own view -> on how google may be thinking about this. So system built on assumptions about another system which we can only speculate on… Anyway great test guys.
Adam:
This is excellent data indeed. There is nothing as convincing as presenting real data to answer a question – and you’ve done that quite clearly.
One interesting question for me though: If you had decided to leave the old domain active, and put a link on every page in that domain to the new domain, do you think the same amount of link juice would be passed? More? Less?
I know the issue of duplicate content clouds this question – so let’s assume you built an entirely new website (not something anyone would probably ever do), would you get just as much juice? I am thinking you might actually get more as the old domain name carries authority and is active. When you put 301s on there, it is now just a pass-through… What do you think?
vancouver web design:
Funny thing is that Google saying that 301 redirect from one url to another will not have negative efect on the SERP not PR. Reality is a bit different, in my experience you always loose some SERP, initially more than you would want, but with time seams that SERP goes back up slowly.
Built Right LLC:
We run a handyman and general contractor services website. Although we ultimately can’t control what people use to link to us, when we find a link to our page that we can dictate where it points, we ALWAYS use the www prefix. Since we do this, the only traffic we concern ourselves with is what flows through there. The 301 redirect then becomes a support tool for better PR and to improve placement on the SERPs, not a deficit.
Richard:
I went to the SMX West show a few weeks ago. While there, the question came up about how much of a 301 redirect is transferred to the new page from the old one. A figure 60% was mentioned, but that figure was qualified as a very vague estimate.
One thing that has been made clear is that cascading 301 redirects is a good way to deteriorate any juice flowing to a page from other (possibly previous versions) pages. So it makes great sense to clean up your site once in awhile instead of putting patches on top of other patches.
Rafay Qureshi:
Great Article But Graphic that you have used in this article are no doubt A+++ = )
And more thanking you for clearing my confusion about the link juice and the backlinks passing on to the new redirected domain . thanksssssssss :)
mian:
It’ll be interesting to see the data of Yahoo and Bing again now. Regarding the old data, have you had anything done to the ‘new’ domain at the point when you 301 redirected it, ie link building etc.
Ron:
Thanks for the infomation Mark, I recently did a 301 redirect with one of my sites and passing the link juice was one of my main concerns. At least now I have an idea of how much juice to expect.
MK Web Designer:
I found very similar results when I transferred one our of customers websites from html to aspx. We used 301 redirects on each and every page to direct the traffic and link juice to the newly named url.
Brian:
Very cool Mark..This has inspired me to do a test of my own. I will follow back up with results.
Jared:
I wonder if it is specific links that they have discredited?
Alex Zorach:
I recently switched my site RateTea from .net to .com, and I noticed no statistically significant decline in traffic from Bing and Yahoo relative to Google. There was a small, temporary decrease in traffic from all search engines, but nothing like the graphs you show here, and the decrease from Google was about the same.
I see this post is from almost two years ago. Things on the internet change rapidly! It could be that Yahoo and Bing got their act together and learned how to properly follow these redirects, or it could be that your one site was an anomaly for some reason. Or it could be that because my domain change was from a .net to a .com, with the same domain, it was treated differently.
Who knows? It’s not really possible to test these things scientifically. I do appreciate your effort in carrying out this test and writing about it, it may not be perfect or scientific but it’s interesting to read about other people’s experiences.
fashionox:
Wow very good information this will help me when I move my website to a new domain.