Change URL or use Canonicals and Redirects?
-
We just completed a conclusive a/b test on a client's landing page. The new page saw a 30% bump in conversions, yay!
Now what?
Option 1: Change the url of the new page to that of the old page, retire the old page.
Option 2: Redirect the old page and anything that was pointing to it to the new page, make the new page the canonical.
I'm afraid of option 1 because I think Google's WTF penalty will be a bit harsher than option 2, but I wanted to sanity check that here.
Any thoughts or experienced advice would be very appreciated!
-
I knew it that sounded like a Google A/B test protocol!
A good rule of thumb is to avoid changing URLs unless it's absolutely necessary. There's a lot going on with that URL in the background that Google knows about....internal and external links as I mentioned above, but also XML sitemaps and usage metrics. You don't want to point them elsewhere and have them re-learn a new URL structure and step through a redirect just to get there.
Google has put more emphasis on UX in the last couple years, so improving the usability of this page, as you've done by A/B testing, is likely to benefit you in the long run.
-
Thanks. We did use Google Experiments, so your advice is very helpful.
Am I crazy in thinking that shifting a completely new page to an old and trusted URL is not going to hurt rankings a bit?
-
Hi,
You definitely want to avoid redirects where possible, so scratch option number 2. Redirection causes you to lose about 10% of the authority that page has built up. Google tends to prefer pages that they have known about for a while.
If you were to do option 2, you'd also have to update all of your internal links to point to the new page, as well as outreach to any external linking sites to have them update.
All you need to do is take the source code for the variation page and make it the source code for the original.
It sounds like you may have used Google Content Experiments. If that's the case, the additional URL created for your variation doesn't need to be excluded from crawls or disallowed, Google knows it's there and there's no other way to get to it other than the code snippet they utilize to send your sample to the variation.
-
Hi,
So if I understand correctly you AB tested with two pages (for example: domain.com/page and domain.com/testpage) and both were indexed by Google? If yes, than option 2 as you mentioned is the best way to go here.
For the future I would recommend to make sure that the testpage is not indexed by Google via robots.txt/meta noindex or use the rel canonical tag. You don’t want the testpage to get organic traffic here to prevent issues.
Got a burning SEO question?
Subscribe to Moz Pro to gain full access to Q&A, answer questions, and ask your own.
Browse Questions
Explore more categories
-
Moz Tools
Chat with the community about the Moz tools.
-
SEO Tactics
Discuss the SEO process with fellow marketers
-
Community
Discuss industry events, jobs, and news!
-
Digital Marketing
Chat about tactics outside of SEO
-
Research & Trends
Dive into research and trends in the search industry.
-
Support
Connect on product support and feature requests.
Related Questions
-
301 redirects for all urls - legal dispute
The website in question is a very high traffic website with substantial credibility in it's subject matter (sorry, can't share more details) that delivers an overwhelming majority of traffic from SEO, much of which is new visitors. A legal dispute has resulted in both parties agreeing to forward a percentage of the total URLs to alternative websites (only 1 website for each party). All URLs for the domain will be forwarded elsewhere. It does not make sense to me that the "sum of the parts" will be as strong once the redirects are implemented but I am looking for feedback. It is fair to say that the alternative domains of each party are no where near as strong as the domain being "parted out." Will the SEO juice be distributed to each domain in full? Will both parties lose out substantially? Feel free to ask for clarifications and I'll do the best I can given the legal parameters. Thanks.
Technical SEO | | ReachMaineAgency0 -
Question on canonicals
hi community let's say i have to 2 e-commerce sites selling the same English books in different currencies - one of the site serves the UK market ( users can purchase in sterling) while another one European markets ( user can purchase in euro). Sites are identical. SEO wise, while the "European" site homepage has a good ranking across major search engines in europe, product pages do not rank very well at all. Since site is a .com too it s hard to push it in local search engines. I would like then to push one of the sites across all search engines,tackling duplicate content etc.Geotargeting would make the rest. I would like to add canonicals tag pointing at the UK version across all EU site product pages, while leaving the EU homepage rank. I have 2 doubts though: is it ok to have canonical tags pointing at an external site. is it ok to have part of a site with canonical tags, while other parts are left ranking?
Technical SEO | | Mrlocicero0 -
Canonical URL on frontpage
I have a site where the CMS system have added a canonical URL on my frontpage, pointing to a subpage on my site. Something like on my domain root.Google is still showing MyDomain.com as the result in the search engines which is good, but can't this approach hurt my ranking? I mean it's basically telling google that my frontpage content is located far down the hierarki, instead of my domain root, which of course have the most authority.
Technical SEO | | EdmondHong87
Something seems to indicate that this could very well be the case, as we lost several placements after moving to this new CMS system a few months ago.0 -
John Mueller says don't use Schema as its not working yet but I get markup conflicts using Google Mark-up
I watched recently John Mueller's Google Webmaster Hangout [DEC 5th]. In hit he mentions to a member not to use Schema.org as it's not working quite yet but to use Google's own mark-up tool 'Structured Data Markup Helper'. Fine this I have done and one of the tags I've used is 'AUTHOR'. However if you use Google's Structured Data Testing Tool in GWMT you get an error saying the following Error: Page contains property "author" which is not part of the schema. Yet this is the tag generated by their own tool. Has anyone experienced this before? and if so what action did you take to rectify it and make it work. As it stands I'm considering just removing this tag altogether. Thanks David cqbsdbunpicv8s76dlddd1e8u4g
Technical SEO | | David-E-Carey0 -
What is the best way to use canonical tag
Hi, i have been researching this since yesterday and have looked at this subject many times before but still cannot get my head around it. i done a report on my site which was very useful, i used http://www.juxseo.com for my site www.in2town.co.uk and it brought me some useful information but part of that info was it was telling me that i should have on my home page a canonical tag which would improve my seo. Now i am using sh404sef for my friendly urls and i am using joomla 3.0 and when i approached the makers of the sh404sef to ask about the tag they said i would need to be careful of using it as it could damage my site and my rankings. i have read lots of information but still do not have a clear understanding behind it. can anyone please explain the best way to use this and should i be using where i may have some sort of duplicate page, any help to understand this would be great.
Technical SEO | | ClaireH-1848860 -
Do I need both canonical meta tags AND 301 redirects?
I implemented a 301 redirect set to the "www" version in the .htaccess (apache server) file and my logs are DOWN 30-40%! I have to be doing something wrong! AddType application/x-httpd-php .html .htm RewriteCond %{HTTP_HOST} ^luckygemstones.com
Technical SEO | | spkcp111
RewriteRule (.*) http://www.luckygemstones.com/$1 [R=301,L] RewriteCond %{THE_REQUEST} ^./index.htm
RewriteRule ^(.)index.htm$ http://www.luckygemstones.com/$1 [R=301,L] IndexIgnore *
ErrorDocument 404 http://www.luckygemstones.com/page-not-found.htm
ErrorDocument 500 http://www.luckygemstones.com/internal-serv-error.htm
ErrorDocument 403 http://www.luckygemstones.com/forbidden-request.htm
ErrorDocument 401 http://www.luckygemstones.com/not-authorized.htm I've also started adding canoncial META's to EACH page: I'm using HMTL 4.0 loose still--1000's of pages--painful to convert to HTML5 so I left the / off the tag so it would validate. Am I doing something wrong? Thanks, Kathleen0 -
Do links that point to an old URL retain value if we have the correct redirects?
I've recently taken over SEO for my company. There are a lot of old links that point to our old URL (www.examplecountry.com changed to (www.examplewhatwedo.com). We have the correct redirects in place and Open Site Explorer shows many of the links pointing to the old site even though I'm inputting the new URL. I just want to put my mind at rest that any value these links have doesn't got lost due to the URL change. Unfortunately a lot of them have the old URL as the anchor text....which I guess will decrease their quality? Thanks!
Technical SEO | | MarbellaSurferDude0 -
Should i use NoIndex, Follow & Rel=Canonical Tag In One Page?
I am having pagination problem with one of my clients site , So I am deciding to use noindex, follow tag for the Page 2,3,4 etc for not to have duplicated content issue, Because obviously SEOMoz Crawl Diagnostics showing me lot of duplicate page contents. And past 2 days i was in constant battle whether to use noindex, follow tag or rel=canonical tag for the Page 2,3,4 and after going through all the Q&A,None of them gives me crystal clear answer. So i thought "Why can't i use 2 of them together in one page"? Because I think (correct me if i am wrong) 1.noindex, follow is old and traditional way to battle with dup contents
Technical SEO | | DigitalJungle
2.rel=canonical is new way to battle with dup contents Reason to use 2 of them together is: Bot finds to the non-canonical page first and looks at the tag nofollow,index and he knows not to index that page,meantime he finds out that canonical url is something something according to the url given in the tag,NO? Help Please???0