If You Are Not Doing Split Testing, You Should Be!
If you haven't given any thought to split testing on your website, you really must. Split testing is the process of running multiple variations of the same page, opt-in form, call-to-action, anything on your website. Ideally these different variations would be run at exactly the same time, with different options presented to different visitors.
You may be the most creative person in the world, or have devoted hours crafting what you think is the perfect squeeze page. Perhaps you think you know your niche better than they know themselves. None of this is relevant at all!
As much as you might think your copy is perfect, let me tell you now.. it's impossible for you to know this for sure unless you have properly tested it.
Sometimes the smallest change can have significant effects. One very famous advertiser, one of the top advertisers in the world in fact, increased sales by 20% on one of his products by changing just one word in a very long advertisement. Just one word. He changed "repair" to "fix" and made 1/5 more sales. There is no way that he could have known this without testing.
When setting up tests it is important to only change one thing at a time. If you set up a test to pick between two different variations of a headline, but at the same time you also change some text on the page, you will never know what the true impact of changing the headline was.
You might have a number of ideas about how you would like to change your headline. Perhaps you want to try a different color, or there is some wording you want to change. Maybe you want to try out some different fonts. Don't get carried away and try to test this all at once.
Conclusive testing relies on there being enough visitors to your site to be able to declare one of your tests a winner. This is to do with the mathematics of comparing two tests with one another. If the number of visitors is too low, then random variations can play a significant part and lead you to making changes that are not supported by the evidence. You may end up in a worse situation than you were originally.
For this reason, if your traffic is low, you will be better off testing one change at a time. Otherwise your tests can take too long to return a conclusive result. Plus it's more fun. You get to make a whole lot of constant improvements. Just knowing that these small changes are going to add up to more sales and hence more profit is highly motivating!
If you have a lot of traffic to your site then you might be able to test several variations at once. There are other advanced statistical techniques that allow you to test several changes at once. The mathematics behind these multivariate tests is daunting and you will definitely want to be using a tool to do this analysis for you.
One really great thing about split testing is that it allows you to feel good about making mistakes. Say you find out that your original headline sucked. Cool, now you know. Or maybe you've always thought that your squeeze page needed to be in a larger font, but the test proves otherwise. Great! Now you know.
That's the point really. With split testing, you KNOW what is best. Without split testing you are doing little more than guessing.
Brett Williams is an internet marketer with a keen interest in A/B and multivariate testing. He recognizes that this can be a daunting subject for many people and has started a blog sharing his early experiences with split testing. The Visual Website Optimizer is a powerful new tool which makes it incredibly easy to get started, even for a total newbie
You can see Brett's experiences on his blog. Check out his first post he made when he just got started here.
You may be the most creative person in the world, or have devoted hours crafting what you think is the perfect squeeze page. Perhaps you think you know your niche better than they know themselves. None of this is relevant at all!
As much as you might think your copy is perfect, let me tell you now.. it's impossible for you to know this for sure unless you have properly tested it.
Sometimes the smallest change can have significant effects. One very famous advertiser, one of the top advertisers in the world in fact, increased sales by 20% on one of his products by changing just one word in a very long advertisement. Just one word. He changed "repair" to "fix" and made 1/5 more sales. There is no way that he could have known this without testing.
When setting up tests it is important to only change one thing at a time. If you set up a test to pick between two different variations of a headline, but at the same time you also change some text on the page, you will never know what the true impact of changing the headline was.
You might have a number of ideas about how you would like to change your headline. Perhaps you want to try a different color, or there is some wording you want to change. Maybe you want to try out some different fonts. Don't get carried away and try to test this all at once.
Conclusive testing relies on there being enough visitors to your site to be able to declare one of your tests a winner. This is to do with the mathematics of comparing two tests with one another. If the number of visitors is too low, then random variations can play a significant part and lead you to making changes that are not supported by the evidence. You may end up in a worse situation than you were originally.
For this reason, if your traffic is low, you will be better off testing one change at a time. Otherwise your tests can take too long to return a conclusive result. Plus it's more fun. You get to make a whole lot of constant improvements. Just knowing that these small changes are going to add up to more sales and hence more profit is highly motivating!
If you have a lot of traffic to your site then you might be able to test several variations at once. There are other advanced statistical techniques that allow you to test several changes at once. The mathematics behind these multivariate tests is daunting and you will definitely want to be using a tool to do this analysis for you.
One really great thing about split testing is that it allows you to feel good about making mistakes. Say you find out that your original headline sucked. Cool, now you know. Or maybe you've always thought that your squeeze page needed to be in a larger font, but the test proves otherwise. Great! Now you know.
That's the point really. With split testing, you KNOW what is best. Without split testing you are doing little more than guessing.
Brett Williams is an internet marketer with a keen interest in A/B and multivariate testing. He recognizes that this can be a daunting subject for many people and has started a blog sharing his early experiences with split testing. The Visual Website Optimizer is a powerful new tool which makes it incredibly easy to get started, even for a total newbie
You can see Brett's experiences on his blog. Check out his first post he made when he just got started here.
0 Response to "If You Are Not Doing Split Testing, You Should Be!"
Post a Comment