Split testing, also called A/B testing, is a testing method where you compare a base sample with another test sample, changing one specific variable inside it. For example, you could have a base landing page, and then you would compare it against a new landing page where you changed the main headline. If you were selling a product on that landing page, you would then compare the conversion rate of page A against page B (by conversion we mean the percentage of the total visitors that ended up making a purchase). There are other testing methods out there (like the multivariate one), but the split tests are easy to implement, easy to interpret, and they can be applied on a wide range of situations. That is why you should always use them.
Let's See Some Examples
Here is a practical example with some numbers. Let's suppose that on top
of your sidebar you have a "Subscribe to my newsletter" section, where
visitors can put their email addresses to subscribe to your newsletter.
First of all you would need to track what is the conversion rate. Just
find out what is your weekly traffic, and how many new subscribers you
get per week. If your weekly traffic is 30,000 visitors and you gain, on
average, 50 new subscribers, then your current conversion rate is 1,66%
(50 divided by 30,000).
Now
you need to change one specific element in that subscription section,
and then track the results for another week or two. Let's suppose that
you changed the background color of that section, making it yellow to
call the attention of visitors. During the next weeks your traffic kept
stable at 30,000 visitors, but you got 65 new subscribers per week on
average. This means that your conversion rate increased to 2,16%, and
you should therefore keep the changes you made. Had the conversion rate
decreased, which is possible, you would just need to revert the changes.
If you have a blog where your goal is to gain as many RSS subscribers as possible, create split tests with the location of the RSS icon, with the message that you use to call the attention of visitors, with the freebies you offer in return and so on. If you have a landing page where you promote an affiliate product, split test with the sales copy, the headline, the testimonials and so on. If you use PPC to drive traffic to your website or products, split test with different keywords and bids and always try to beat the top performing ad group.
This is pretty much the process you'll need to repeat over and over again, always trying to beat the top performing combination of factors. Obviously you should not split test every single element on your site, else you would not have time for anything else. But it is very important to take the time to split test the important sections ( i.e., landing pages that sell your products, pages where your you get email subscribers and so on).
No comments:
Post a Comment