More E-Marketing Tips |
How can you increase profit with the same amount of traffic? TEST new things. Testing and tracking is crucially important for the profitability of your site. You can literally test everything on your sales page , but there are some basic things to test first. • Headline • Header Graphic • Introduction • Price • Call to action • P.S. Messages
The simplest way to test is an a/b split test. Here you create two pages that are identical apart from one element , for example the Headline. The split test will rotate between the two versions, which ever gets the most conversions in the end wins.
The highest performing version is the “control” , you attempt to beat the control with each new variation. Once you’re satisfied that there is no more improvement, move on to the next element.
You can set up your split test by signing up for an Adwords account and use Google Analytics Content Experiments to do your testing. Create the versions of your page and add the tracking codes provided to the pages you want to test.
It’s important to know if a test is valid. With the Google program it will finish by itself once it obtained a “valid” result, but if you are using another program you would want to make sure that your results are statistically significant.
Warning! Maths ahead. I will try to make this as simple as possible. With any result there is going to be statistical variation. Even with two identical pages it is highly unlikely to get a 50/50 split. You will more likely get a 53/47 split , that doesn’t mean the one is better than the other because they are identical.
Statistical significance can be worked out as follows. If the difference between conversions between A and B is greater than the square root of the total conversions then the result is statistically significant. Okay I know that hurt, but let’s clarify this by use of an example.
Example 1: A/B test runs until a 100 conversions have been obtained.
A: 56 Conversions B: 44 Conversions N=100 that is the total number of conversions.
Difference between A and B = 56-44 = 12 Square root of conversions is 100=10 The difference (12) is greater than the square root (10) so the test is valid.
Example 2: A/B test runs until a 100 conversions have been obtained.
A: 54 Conversions B: 46 Conversions N=100 that is the total number of conversions.
Difference between A and B = 54-46 = 8 Square root of conversions is 100=10
The difference is less than the square root so the test is not valid in this case.
Google will do this for you so don’t break your head over it, I just want you to know how to do it by yourself. Just remember to let the test run for long enough, if after a 100 conversions you don’t get a valid result, stop and test something else.
There is also a more advanced way of testing your web pages, where you can test a bunch of elements all at once. This is called multivariate testing and it works best for sites with quite a lot of traffic (1000 per week or more ideally) . Multivariate testing runs multiple combinations against each other and find the best one. Testing 5 different elements will give you 5x5 =25 combinations. Testing 6 elements will give you 6x6 =36 combinations. This allows you to test faster than with A/B split tests. Google has good tutorials to guide you through the whole process, don’t worry , they do all the calculations for you.
Apart from testing your web pages you must also test your ads. Facebook allows you to make six different variations so you can see which ad performs the best and then you can scale up the ones that performs best.
Essentially you want to test everything you can think of: • Squeeze page • Sales page • Opt-in Form • Autorepsponder messages • Thank You page offers
Effective testing will not only improve your profit per click, but it will also guide you on what will work on other sites that you might have.
P.S. For more on testing you can read the book “Convert Every Click” by Benji Rabhan. He has some very good testing techniques in there.
Related Articles -
split, test, your, squeeze, page,