What is A/B testing and why do you need it for your business?
When you are building your great new web site for your business, you of course get your best graphic designers and interface wizards on the case and you (hopefully) end up with a fantastic looking site which impresses not just you and your mum but also all your real visitors and then you launch it online with much fanfare.
It’s a great looking site and you have a great product so of course when the traffic starts rolling in, a respectable percentage of it converts into sales and so you assume all is good. But could it be better? Almost certainly!
How do you know your site is converting the best it possibly can? Many times what you think will convert the best in fact does not and you may never know what you are missing. This is where AB Testing comes in and it is critical that you do it. Not doing A/B testing at all is a grave error to which a great many fall victim, on a scale similar to Julia Roberts in the movie “Pretty Woman” being practically booted out of a posh clothes store in Beverly Hills by snobby sales staff – Big Mistake, Huge!
A/B Testing (also often called split testing although that also includes other types) is simply the side by side comparison of two versions of your site (or usually a specific page on your site such as the index page or a purpose made landing page) to see which one converts the best. You have your original version A and you change something, for example some text, or a graphic or a button or colours or even the whole design and that is version B. Then you send equal amounts of traffic to both over a period of time and you will soon see which one converts best, so you go with that one. Then you make another change, another version and test again and repeat that process over and over making changes each time until you have a highly optimized final version which will convert much better and so you make much more money and everyone is happy.
Low Hanging Fruit
When starting out it is usually better to focus on big changes – make two very different landing pages for your tests. Making small changes can sometimes make a big difference but in most cases the improvements will not be so dramatic and if you don’t have huge amounts of traffic you could be spending months on testing these tiny variations without making any significant discoveries. Big changes are more likely to have a big effect which you will notice very quickly and so have the capacity to dramatically improve your conversions much sooner which for a startup or SMB can mean you get to stay in business.
It is commonly said that one should alter a single thing at a time when AB testing so you know exactly which modification caused the change in conversion rates, but realistically when you are starting out, do you really care exactly why one page converts at 10% while another converts at 1%? You’re making 10x as many sales so take the money and run!
You can always refine your top converting pages over time with progressively smaller changes and further A/B testing and eventually you may be reduced to testing different colour signup buttons or fonts to get that last fraction of a percent increase in conversions – that’s a good place to be but it’s not a good place to start!
In the future, when you have more traffic and more resources you can move on to more advanced testing, such as Multivariate, and start going for those harder to find optimizations. Until then – go for the low hanging fruit!
When performing such testing it is of utmost importance to have tracking and analytics in place so that you can properly measure and act on these modifications as you go along. Womplify in fact provides tracking capability which can be easily added to your landing pages for just such a purpose.
There is another variation of split testing called multivariate testing , as briefly mentioned above. With this method you can create multiple variations to test at the same time, instead of just A versus B each time. This is of course a faster method to test many changes, particularly small changes, but comes with added complexity to implement and analyze the data correctly and typically requires considerably more traffic to produce significant results.
A/B Testing Tools
The technical implementation can be a little complicated if done in-house with custom code, so many companies will outsource split testing to an agency or conversion consultant, however there are a number of online tools available which you can use to simplify the process.
Google provides “Google Content Experiments” which replaces google Website Optimizer and is included in their Analytics service for free. There is also a WordPress plugin available. Content Experiments is a decent tool and good enough for many people, however there are other good tools available, such as..
Visual Website Optimizer – prices start from $9 per month but is rather limited so you would want to go for the $49 plan for best results.
AB Tasty – prices for this one start from €29/month.
Whichever method or tool you choose, it is important to take the necessary time to do your AB or Multivariate testing thoroughly so you can be assured of having the highest conversion rate possible from your web site, otherwise you could be needlessly leaving a LOT of money on the table!
What are your experiences with split testing? Which is your favourite tool or testing method?