Testing and optimizing your campaigns and website is crucial for affiliate marketing success. Dedicated A/B testing distinguishes a beginner affiliate marketer from an expert. In this guide you’ll learn how and what to test as an affiliate marketer. Let’s boost your affiliate conversions!
This guide is for you if you want to start A/B testing for:
- Content that includes affiliate promotions
- Web pages and your website interface
- Affiliate email campaigns
What is A/B testing in marketing
A/B testing, also called split testing, is the process of trying 1 or more different factors to see which one gets better results. A/B testing is common in many professional industries and also plays an important role in marketing. A/B testing is frequently done to test email flows, websites, and other digital marketing activities.
Why do we need A/B testing in affiliate marketing?
In affiliate marketing, A/B testing is needed to evaluate your conversion funnels. Whether you’re just starting out or you’re a seasoned pro, a data-driven attitude will help your business forward. Guess work simply doesn’t cut it. You will be surprised by the results you’ll obtain through experimenting!
A/B testing for affiliate marketing:
- Provides insights into how your audience responds to changes on your website or in your campaigns
- Is data-driven and eliminates guesswork
- Helps identify issues that influence the user experience on your site
Well done A/B testing directly impacts your conversions and affiliate commissions.
How to set up A/B testing as an affiliate marketer?
Perhaps you have a multitude of websites and campaigns running to get affiliate referrals. Each aspect of your campaigns can be tested to find elements that give a better conversion rate.
A/B testing requires a so-called “control version“. This is your website, email, or piece of content in which you don’t change anything. It also involves another version where you change a single variable. Both versions are shown to similar sized audiences. After this you analyze which version performs better over a set amount of time. Next, you implement the version with the best results.
The process of A/B testing for your affiliate site and campaigns looks like this:
Let’s go over this process step by step.
1. Decide on your testing goals
Formulating your goals is essential for split testing. Goals are necessary to evaluate your results. Ultimately all goals are linked to the success of your affiliate business, but we need to break them into pieces to create a good test environment.
Some examples of goals used in A/B testing:
Your content needs to match your audience and niche. But that’s not as easy as it sounds. In most cases, it should be informative and interesting, perhaps even entertaining depending on your niche. You also need to come up with a great headline and appealing images. Consider what type of content is important for your audience and topic and base your content goals on this. Split testing different elements of your content can help you create content that converts.
- Lower your bounce rate with the help of engaging content
- Inform your audience so they can make the right purchase decision
Your product review has the goal of convincing readers to buy a product
Different pages have different goals. Formulate your goals for specific pages so you have a direction for your experiments.
- Your landing page serves as a bridge page to send people to the affiliate offer
- The goal of your contact page is to get in touch with your audience
- Your product review page has the goal of sending referrals through your affiliate link
Your split testing goals need to be aligned with the goals of the webpage you’re targeting.
Website interface and speed
The level of engagement you get from your visitors highly depends on how easily they can navigate your website. Perform a technical audit of your website and set goals to improve your website speed and solve all technical issues. A low website speed, for example, can affect your affiliate conversions.
You can test almost anything on your website. And there are many optimization factors hidden from our own sight. Follow your users behavior in Google Analytics and you will understand where your conversion pain points are.
Email campaign goals
If you want to boost your affiliate referrals through email campaigns, create specific goals that help you achieve higher affiliate commissions.
- Increase your on-page email sign-ups
- Boost your email open rate
Improve the click-through-rate of your affiliate promotions
The key is to break down your goals into smaller pieces, so you can test and optimize for affiliate success.
2. Formulate your hypothesis and create variables
There is an overwhelming amount of possible factors to test. Formulate a hypothesis based on data and previous observations.
Website data from Google Analytics can help you get insight into your users behavior. A tool like Supermetrics can help pull all your marketing data in one place, so you can evaluate campaigns with testing potential. Collect research data that allows you to evaluate your assumptions for creating more conversions.
A good hypothesis explains what needs to be changed and how it helps reach your goals. The A/B test can then either prove or disprove your hypothesis.
Don’t be tempted to test multiple factors in the same test. Changing 1 variable is the best way to assess the difference in results. Small changes can have a big impact on your click-through-rate and affiliate commissions.
Decide on your control and test variable. If you’re experimenting with your website, this means the control factor is your website without any changes, and your change is the test variable. The variables to test should be aligned with your goals.
3. Set your sample size & statistical significance
Proper split testing can only be done if the test and control group is big enough. A test group that’s too small may lead to wrong conclusions.
Your sample size always needs to be big enough, even if your affiliate website does not have much traffic yet. Without knowing your sample size, you risk stopping the experiment too early. The larger your sample size, the less margin of error in your conclusions.
To calculate the right sample size for your test, you need to determine the level of significance. There are several sample size calculators online that can help calculate the sample size for your test. You can use this sample size calculator from Optimizely, for example.
Decide how significant your results need to be to justify choosing one factor over the other. The significance level defines the distance the sample mean must be from the null hypothesis in order to be considered statistically significant. A/B testing can only take place if you keep in mind the concept of statistical significance.
Another concept is the so-called “confidence level“. The higher the confidence level, the more sure you are about your results. As an affiliate marketer, you want to be certain about your results as they directly impact your income. Reaching your statistical significance level excludes chance as a reason for your results.
As you can see in the above image 95% (the confidence level) of the values are within a normal distribution and 5% is the significance level. This means there is less than 5% probability of outliers that are beyond 2 standard deviations from the population mean.
You should always aim for a confidence level of at least 95%. This 95% means that if you would repeat the test, 95% of the results would be the same.
As mentioned before, the level of statistical significance has an impact on your sample size. You can read more about sample sizes for A/B testing and statistical significance in this article by Conversion Sciences.
The lower your traffic, the longer you will have to run the test in order to obtain enough data. Keep in mind that other factors can impact your traffic, such as seasonality. If your niche is tied to seasonal products, you want to be careful when to run your test. It’s also a good idea to test your variables on different traffic sources and user devices.
4. Run and analyze your experiments
Depending on your situation and the amount of variables, you need to pick the right testing method. Beside the straightforward A/B testing method, you can consider:
Split URL testing
A/B testing is done when front-end variables are measured, while split URL testing is done when there is a big change to be made in the design, and you want to keep your existing web page design available.
This test allows you to check multiple versions of a webpage which are hosted on different URLs. Your traffic is then split between the different URLs. The page with the best conversion rate wins.
The difference between split URL testing and regular A/B testing, is that for split URL testing, the variables are hosted on different URLs.
Multivariate testing is needed when testing multiple variables at the same time. The test helps you find the best possible combination of factors to reach your goals and can point out which variable has the biggest impact on your conversions.
Multi page testing, also known as funnel testing, lets you test changes on multiple pages at the same time. Where A/B testing lets you see the effect of a single change on a single page, multi-page compares multiple pages.
Many of the principles mentioned in this guide work on all these test methods. However, these tests are more complicated than regular A/B tests and it’s best to first fully understand A/B testing before jumping into other types of experiments.
Testing tools on the market
There are several A/B testing tools on the market. Google Analytics, for example, lets you A/B test multiple versions of a webpage and compare their performance. For this experiment, Google Analytics uses random samples for users.
Run your test
Run your test for the appropriate amount of time to collect enough data. You can calculate the running time for your experiment with this calculator at VWO.
After the test, check all important factors to screen for possible errors such as:
- Significance level
- Sample size
- Type of visitors by traffic source and device
- Amount of affiliate referrals
If everything went well and you have collected enough data, it’s time to analyze your results! Each variable tested will either win or lose compared to the control variable.
Analyse your results
If your test was set up well, analyzing the results is straightforward. You can analyze your results for example by percentage increase or impact on other metrics.
Analyze if your test variable is responsible for the desired goal. In other words:
Did your hypothesis turn out to be true?
If the answer is yes and your variable turns out to increase your conversion rate, implement the winning variable. If you need help tracking your affiliate conversions across different channels, check out these 5 dashboards to track your affiliate sales.
If your variable did not have the desired outcome, evaluate reasons why the test failed. It’s important to use these conclusions in future experiments. Go through your hypothesis and reevaluate your research data.
Start over and repeat the whole process to optimize your affiliate activities.
A/B testing examples for affiliate marketers
Let’s go over a few A/B testing examples that can boost your affiliate conversions.
1. Split testing landing pages
If you want to send people to your affiliate offers with the use of a bridge page, you can start split testing your landing pages. This helps you determine which design and copy works best to convert your audience.
Send 50% of traffic from the same traffic source to 1 landing page and 50% of traffic to another landing page. Both landing pages lead to the same offer. After this evaluate your landing pages and keep the best converting one. Split testing landing pages can give you a quick answer to which page works best.
2. Direct linking vs using a landing page (bridge page)
You might be unsure whether an offer converts best by sending direct traffic or with the help of a bridge page. There are trends for niches and their preferred traffic sources and funnels, but in case of doubt, an A/B test is easy to execute.
Use the same promotion and traffic source to send 50% of traffic directly to the offer and 50% to the bridge page. Evaluate which method leads to more conversions and continue with the winning approach.
3. Test your headlines
Your headline is the first impression of your content piece. A good headline makes your readers feel curious to click and find out more.
Who will benefit from this promotion?
Is this a new strategy or product to combat a common problem in your niche?
Highlight how this content will help your readers. Try to describe what is unique and different about your product. These are key points for your headline because they help create curiosity. Brainstorm a few headlines.
Next, choose your favorite two headlines and show each headline to 50% of your traffic. Depending on your traffic source, you can do this in several ways. For paid promotion, create ads with different headlines leading to the same content. If you get traffic to your blog using email campaigns, set up half of your emails with one headline and half with the other.
Analyze the results. Which headline performed the best? Keep the headline with the most clicks. The bigger goal of trying out different headlines is to find the type of language your audience responds to. Once you figure this out, you can tailor all your headlines towards this.
Other content elements to experiment with:
- The calls to action in your article
- The length of your content
- The position of your suggested posts
- The position of your affiliate promotions within the content
4. Try out different images
Images matter as much as headlines. They have to relate to your content, bring curiosity and inspiration, or evoke other emotional responses.
People are naturally drawn to images that include:
- Human faces: our eyes stare longer at images with human faces
- Portals: windows, tunnels and stairways do great in capturing your readers attention
Pick two images to test within the same article. If you want to feature an image of your affiliate product, pick two appealing images of the same product. Always pick high-quality images. Split your traffic and test the images. Analyze the results and keep the best converting images.
5. Test your ads
In case your affiliate program or network allows paid advertising, you can experiment with your ads. There is a lot you can test and optimize for your ads. Some variables to test are :
- Advertising platform
- Headline of the ad
- Body of the ad
- Device the ad is displayed on
- Calls to action and add extensions
6. Experiment with you email copy
Your email copy, or the body of email, is there to persuade your readers to click on your affiliate link. Experiment with altering your wordings in certain parts of your email. You can, for instance, use words that trigger urgency and a fear of missing out for the promotion. Also, try altering the font, length, and placement of sentences to make them stand out.
Other elements worth experimenting with are email design, layout, and calls to action. Send each of your two email variants to 50% of your email list and track your results. Besides the open rate, also keep an eye on the click-through-rate, and the amount of referrals you actually end up with.
7. Try out different subject lines
Another important email experiment is to test your subject line. Create two appealing subject lines and track the open rate of the email. If you have a large email list, send out version A and B to a small percentage of your subscribers. Select your recipients in the test group randomly.
Once you find out which subject line has a higher click-through-rate, send out the winning version to the rest of your email list. After the test, try to find patterns of successful subject lines.
Improving your affiliate activities is an ongoing process. If you want to learn more about driving affiliate conversions, check out these affiliate marketing strategies to boost your conversions.
About Hetty Korsten
Hetty Korsten is a Partner Marketing Manager at Supermetrics. She has worked for fast-growing SaaS startups in Copenhagen and Helsinki. Currently she’s growing the in-house partner program at Supermetrics. Feel free to connect with her on LinkedIn.