Conversion optimization with Optimizely, impact on SEO

[Robertnweil1]

I am looking into signing up with www.optimizely.com for conversion testing.  They put scripting on my site which will then redirect half of my site visitors to an alternate home page.  The site I want to test on is ranking quite well with Google and I do not want to hurt my rankings.

And with this set up,what Google will think is my home page is not and so I am essentially sending visitors to a different page than Google reads as my home page in source code.

So, my concern is whether this will have a negative impact on my SEO rankings to redirect 50% of site visitors to a different page using this testing tool?

I would use Google Web Site Optimizer, but many of my sites are in Wordpress and it seems that Wordpress and web site optimizer are not so compatible.

Advice would be appreciated.

Thank you,

Robert

[AdoptionHelp]

My understanding is that the same URL is used for both test and control versions of your home page, meaning its not actually redirecting. Instead the changes are inserted using the javascript code dynamically. I'm not sure whether Google ignores the javascript change or incorporates it. Either way your test changes should be done with both SEO and user experience in mind, so don't run tests that would hurt your SEO and you'll be golden.

[Robertnweil1]

Hi Ryan, thank you for response.  There are two ways to test with them.  The type of testing I would want to do would apparently (according to sales guy) require an alternative landing/home page.

[AdoptionHelp]

Then you'll want to ask him if that necessitates a separate URL. If it does, you should rel=canonical the test page to point at the original home page, and noindex the test page to avoid any duplicate content situations.

You may also want to run this by the optimizely guys to see what their thoughts are, since they'll know exactly how their technology is altering your site.

[dsiroker]

Hi Robert,

This is a great question. You are right that there are a few ways to run an experiment with Optimizely. The default and easiest way to get up and running is to run an experiment on the same URL without redirecting to another page. Alternatively you can redirect to another URL. Customers commonly use both techniques and neither will negatively impact SEO.

Here is a more detailed explanation of how Optimizely affects SEO: http://support.optimizely.com/kb/advanced/does-optimizely-affect-my-search-ranking-or-seo

Here is Google's stance on this: http://www.google.com/support/websiteoptimizer/bin/answer.py?hl=en&answer=63382

Hope that helps answer your question!

Thanks,

Dan Siroker

Co-Founder & CEO, Optimizely

[kennyrowe]

Yes, I've used Optimizely, and Ryan is completely correct in that  you only create one page, just add javascript to make it different views for different visitors.

[Robertnweil1]

Thank you for all of your responses.  Dan Siroker, nice to see you responding- too thank you.  I suppose I am left with that if I do use/need the function in Optimizely where I need to re-route to new url is to potentially use rel-canonical and block the second page in robot.txt.   I feel that for basic landing page testing, I would not re-route and Optimizely seems like a great and easy solution and I will use it.  For tests which would require re-routing, I am thinking perhaps Google optimizer is better because I am confident they will not penalize for re-route when they are the ones doing it.  Dan, if you feel different, I'd be happy to get on phone to chat.  Thank you all.  Best,  Robert

[diywm]

I realise this response is almost a year old, but I've only just found it. I am about to start using Optimizely, but am concerned about affecting my ranking especially now that we have a new Penguin update due any time now.

The nice guy at Optimizely told me today (during the webinar for newbies) that you could use Optimizely to permanently direct different sorts of visitors to different content. For example, he pointed out that people who today run IE 6 are very different from people who run IE 9. And depending on your market it might make sense to give those two groups different messaging.  But Google specifically say don't deliver different content based on user-agent.

http://googlewebmastercentral.blogspot.co.uk/2012/08/website-testing-google-search.html  And on the last point in that article, how long is too long?