One of the biggest advantages of making money from home with affiliate marketing compared to a regular job is that you can constantly improve the rate at which you earn using tracking, testing and optimisation techniques. Not only is this a perk of the job, but having growth is necessary to just maintain a level of income over time since offers die off, ad networks go down, and the competition increases on a regular basis. While most affiliate marketers have a general idea that they should be testing, a lot of the people who actually put in the effort to test don’t do it properly and don’t have the right mindset about it. Being better at testing will give you a large advantage over the competition.
The Correct Mindset: Optimisation Never Ends
There are many variables that determine the profitability of an affiliate marketing campaign. The proper optimisation mindset is based on the fact that there is no such thing as a perfect affiliate marketing campaign because the optimal level for each variable changes over time. As an example of this idea in action, changes in variables can be seen over time if you leave a campaign alone without changing anything at all because of the volatile nature of market conditions and currency inflation and deflation. Because the optimal value for each variable changes over time, your attempts to improve your campaigns will never end. If you did happen to ever come across the “perfect” affiliate marketing campaign where all of the variables were perfectly optimised, it would only last for a moment since the optimal values for each variable would change just a few minutes later and the campaign would need to be tested to find improvements.
Basic Split Testing
The concept of A/B testing, also known as split testing, is pretty straight-forward. You choose a variable that you want to optimise, and you choose two values for that variable which are typically referred to as value A and value B. You have your campaign randomly choose value A or value B for a period of time so that you’ll get a sample of results to analyse. After you’ve tracked the results for some significant period of time, you choose which between value A and value B was the best-performing option, and you call this value the winner. The loser is then replaced with another value for the variable being optimised, and the process repeats itself. Over time, this method allows you to consistently improve on your best results by finding more efficient values for each variable.
Statistically Significant Samples
When you’re choosing winners and losers from the results of your tests, you have to be careful that you use statistically significant sample sizes. In general, this means that the larger your sample and the larger the difference is between the performance of value A and value B, the more likely you are to be choosing the true winner of the contest. For example, suppose that you’re split testing some advertisement to see how often it gets clicked using two different colours. If out of a sample size of 10 impressions value A received three clicks and value B received two clicks, that’s not a statistically significant result since it could easily be giving you the incorrect winner. However, if the sample size was 500 with value A receiving 110 clicks and value B receiving just 65 clicks, then you can be pretty sure that value A is the true winner because the results are statistically significant.
Strategies for Discovering New Test Values
One of the hardest parts of split testing is figuring out new values to try for some variable you’re trying to optimise. Sometimes the available options are obvious, like when you’re given a set list of colours you can use for a contextual advertising block. However, most of your testing will be of variables that have a subjective element, like the header text for a page. In this kind of situation, it can be very difficult to come up with ideas for new test values since there is a certain amount of creativity involved. One popular option is to borrow ideas from your competition, or at least using the competition as inspiration. However, a more universal and lasting way to dig up new ideas is to think in terms of what reaction you want the reader to have, and see which emotional appeals perform the best instead of focusing on the specific text being used.