As an affiliate marketing expert, you are constantly working to improve your digital channels. An A/B split test helps you to validate whether what you change on your website actually has a positive effect.
Table of contents
Oftentimes when doing an A/B split test on complex websites, you need a little help from a developer.
As an affiliate marketing professional, however, you can easily set up an A / B test yourself.
In this article, I will share 5 practical examples and walk you through the process.
Before you start, write it down on paper.
Before you get started, put the experiment on paper.
What are you going to test, and why?
Keep in a roadmap of the experiments you set up. That way, you keep track of which tests you have done and which are still in the offing.
That way, you can also determine priorities. Which A / B test do you set up first? Describe the following parts per experiment:
Name of the test
- What is the name of the experiment? What are you trying to test out? This is often the same name that you give in the tooling with which you set up the test, such as ‘Google Optimize.’
- What is the hypothesis for the split test? Base your hypothesis on data so that you can also make the experiment measurable. After all, you can’t ignore the marketing data, right?
- Briefly describe how you organize the test. What are you going to change about a web page?
- Is the test for everyone or only if visitors are using a particular device? For example, mobile devices.
- How will you determine whether the test was a success? Think of KPIs such as clickthrough, conversion, leads, and clicks on a specific element such as a phone number.
Watch out for…
- Are there any risks in this test? Your split test page will soon influence the behavior of your visitor. Preferably in a positive way, of course. But if there is a risk, write it down so that you will bear in mind possible problems. Describe what you will do if the risk arises (for example, affiliate sales will drop – how will you fix it?). It may mean that you have to end the A/B split test early.
Business impact on your affiliate marketing campaign
- How big is the business impact of your split test? On a scale of 1 to 10 (little, much), this is a useful mechanism for determining which test has a priority.
Share your experiment with colleagues. This way, everyone is aware that you are doing an A / B test. Perhaps more people can participate and give you direct, first-hand feedback.
1st Example. Adding a (secondary) call-to-action
New registrations of any affiliate marketing campaign are an important KPI for everyone who is being paid for leads. To increase the number of leads (registrations), you can add an extra call-to-action on your website or page.
This secondary call-to-action allows visitors to register in case the primary call-to-action will not catch their attention.
Side note: If your call to action button is left unnoticed, it’s a problem. Instead of split testing, roll out your sleeves and make changes to your design.
When I tried A/B split test on my website, the addition of a secondary call-to-action on the webpage resulted in an increase in conversion rate on new registrations at +206%. The conversion rate on the primary call-to-action also increased by +21%.
2nd Example. Order of the content
In another experiment for the same website, I have changed the order of content blocks on the homepage. I noticed that the theme block of content consistently had the highest clickthrough.
Was it because the content was appealing, or because it was in the middle of the page? By re-arranging the order, I wanted to validate whether a position affects the clickthrough.
The position of a theme did influence clickthrough.
I have tested all possible positions. One alternative came back as the winner, with the order on all content blocks resulting in a higher clickthrough.
I tested that segment again, but this time on conversion as a success metric. And guess what? This modification ensures a higher clickthrough and an increase of +10% in conversion rate.
3rd Example. Position of content
This is a famous marketing case. From an analysis at Philips, I noticed that visitors did not fully understand how a subscription to toothbrush cups worked. Although it is mentioned on the web page. In this A / B test, they adjusted the order of content on the web page.
The information on how a subscription works have been moved up.
I believed that informing the visitors earlier about how a subscription works thus increased conversion.
By bringing up the content with information about the subscription, the conversion rate decreased by -15.35%: surprise, surprise.
While I still believe it is a good idea to inform visitors about the subscription, this was not the right position to do so to increase conversions.
If you are in the affiliate marketing business, your conversions are often either leads or sales. Subscriptions are harder to sell, so you might consider when you build a website to promote an affiliate subscription.
4th Example. Split URL on product page
Do you want to test a completely different experience? Then a split URL test offers a solution.
With this strategy, you direct part of your visitors to a completely different web page. You can do it for selling different products based on location or device, or simply test the design for various affiliate marketing offers.
One of the split testing pages resulted in an increase of +110% in visitors who put the product in the shopping cart. A great result!
At the same time, such an experiment is inconvenient to set up.
Because what was the reason that caused more visitors to put the product in the shopping cart? Was that the product box? The way the copy was presented? A combination of both? Since we don’t know WHY the second page had more people putting their products into the shopping basket, we cannot build on that strategy in the long run.
However, we can say which affiliate page converted better.
5th Example. Display of prices
Travelers who travel in their car often purchase parking spaces online. However, I noticed that many travelers left negative feedback after their reservation.
The price was advertised on the affiliate product page, but that often deviated (strongly) from travelers’ price for their chosen period.
In this experiment, I removed the prices from the web page.
The idea was to validate whether negative feedback decreased as a result, but also whether this had an effect on the conversion rate since we were using affiliate links that were directed through our affiliate software, and we could analyze all the benchmarks of our traffic.
By not showing the prices on the product page, negative feedback decreased by no less than -52%.
And, the conversion rate on the parking products? That even increased a little by +0.21%. In other words, the conversion remained the same, and the negative feedback decreased distinctly.
Use an A / B split test, but not just because you can.
Anyone can actually set up an affiliate marketing test to compare different performing campaigns, offers, or landing pages. Just make sure it is clear to you and colleagues why you are setting up that test.
Avoid experimenting for the sake of experimenting. Just because it’s possible. Data is your best friend here.
With a substantiated hypothesis and a solid roadmap, you can optimize your affiliate marketing channels in no time. Especially if your experiments are based on data and not on gut feeling.