Most of us want to make from our websites, whether it is more newsletter signups, more use of social media button, more leads or more sales. Most of us go about this by trying to increase traffic to our websites whilst often ignoring or spending far too little time on optimising our sites to increase conversions. If you can test and tweak to increase your conversions (by conversions we’re talking about a measurable actions such as a newsletter sign-up or a sale) to produce a 20% increase then this can be just as valuable as a 20% increase in traffic and can be much cheaper and easier to achieve.
What Is A/B Testing and Multivariate Testing?
All types of testing require a measurable action (such as a sale or lead) to be tested. A/B testing does what the name suggest, it test version A of a web page against version B. Let’s say you have a page that you are trying to get more newsletter sign-ups from, version A and B might have a different sized sign-up buttons, you would send half the traffic to A and half to B and measure which one had the most newsletter sign-ups (the highest conversions rate). Multivariate testing differs from A/B testing in that you can test multiple pages or multiple elements at the same time to see which combinations produce the best results. For example which combination of colour and size of ‘sign up now’ button works best alongside where it’s placed on the page and what image is used etc. Whilst there are free resources for multivariate testing (more on this in the next section of this post), for big commercial sites it’s worth considering getting some help from a company specialising in this area such as Maxymiser who offer multivariate testing and can help with the design, implementation and analysis of the tests and have experience of what works well and where the best results can be achieved.
How To Run Conversion Optimisation Testing For Free On Your Site Or Blog
If you’ve never run any form of testing before A/B testing is a good place to start, it’s simpler and can give fast results for some quick wins. If you are A/B testing the first thing to decide is what action you want to measure. Secondly you will need to come up with a version A and B of your pages to test. Thanks to Google’s Website Optimizer A/B testing (and multivariate tests), are free to run. There are some useful video tutorials which can be found in the Education section Google’s Website Optimiser to get you started. It’s actually much simpler than you might think so don’t be put off by thinking it’s too technical and given’s it’s free too it could be the move the takes your site to the next level. If you’ve given A/B or Multivariate testing a go and have seen some good results we’d love to hear about it and hopefully some success stories will encourage others to give it a go please leave a comment and share your experience.
(Image credit Daniel Waisberg)
June 2, 2011 at 4:29 pm
Interesting article, I’m going to try some AB testing for different social media buttons on my site, I’ve been trying to decide between the scolling ones (like on Techsling) and fixed ones at the top and bottom of the page so hopefullly this will help put some science behind my decision; I’ll report back and let everyone know how I get on!
August 29, 2011 at 10:58 pm
So how did it go? Did it work out?
Great article, I didn’t know google offered this. I will try it with my next site.
June 3, 2011 at 8:06 am
Your tips will definitely help me, thank you..
June 7, 2011 at 2:42 pm
thank you very much for posting the website multivariate testing options especialyy lots of thanks for the how to run conversion optimisation testing For Free. Anyone loves free cheese and I am not an exclusion, besides I would really like to see how healthy is my website.
June 30, 2011 at 4:25 pm
Note that this kind of testing can not tell how the different variants perform as far as the search result position is concerned. So unless one tests here exclusively with regard to AdWords conversion, any changes of the page design on the basis of such testing could actually backfire.
Teake van der meer
September 18, 2011 at 5:44 am
I’ve used A/B testing many times.
It is one of the basic things they teach you in business-school.
What I didn’t know is that Google makes it that easy.
I always do it manually.
And multi-variate testing is even better. that way you can test even the smallest details.
Really interesting to know!
February 29, 2012 at 2:03 pm
It is a very powerful tool, but depending on what you select as your conversion goal, you often need a lot of traffic to provide a statistically significant improvement – which can mean leaving one running for quite a while.
It’s also worth doing user testing on both versions in case there is a group of users you are excluding, although overall conversion rates may increase.
March 29, 2012 at 11:41 am
I highly appreciate your articles. It has got such an exceptional view. I would be very much interested to read such more informative draft from you ahead.
Such an effort will really be helpful for readers like me. Keep up the good work.