When you design an ad campaign, many of the variables you choose to use may seem arbitrary. However, even minor changes such as color, font, or image selection can make a substantial difference in converting visitors or followers to customers. Before you commit to a single ad concept or pitch wording, test your advertising to see just how effective it can be.
What is Split Testing?
Split testing or A/B testing is the process of evaluating two versions of something, which are the same except for one variable, to see which version performs better. Variations often involve visual differences, such as changes in wording, color choices, and image selection. You can also split test CTAs.
7 Types of Split Testing
Now that you understand conceptually what split testing is in theory, let’s look at how it can be used in practice. You will see that you can use split testing just about anywhere – and you should. Companies with the most effective marketing do not get there by accident. They test and retest until they develop the perfect formula for appealing to their customers.
1. Different Calls to Action
One common thing for which to use A/B testing is on different calls to action. Many marketers test which word series works best for their customers. Something as simple as “click here” versus “learn more” could be significant. You should also test smaller variations, such as “click there” vs “click here.” This helps you understand how your customers like to receive direction and the language they prefer you use.
2. Content Choices
Also, consider using split testing to evaluate content choices. Think about an accounting firm. It could post two articles on a new small business tax: “X Things You Need to Know About the New Tax” and “Understanding the New Tax Guidelines.” Then, that firm could measure which of those approaches get the most clicks. Some readers prefer a certain type of content. Additionally, you may find that using a certain structure gets you more clicks.
3. Copy Formatting
Copy formatting also works well with split testing. You could look at whether the length of the paragraphs, the use of headers, the overall length of the piece, or the use of bullet points makes a difference in how effective the piece is in attracting leads, visitors, or more clicks on your site. Just remember to test only one variable at a time!
4. Visual Variations
Look also at possible visual variations. Many companies will use A/B testing to decide which font to use, the best color palette, the right graphic, the location of widgets on the website, the size of the headline – if you can see it, you can test it. Marketing is very visual and people do respond more to certain images. By testing visual variations, you can see what appeals to your customers.
5. Sales Funnels
A/B testing can be much more involved too. It can go deeper than looking at how many people clicked on something. Many companies use split testing to evaluate their sales funnels and map the customer journey.
Split testing has a role in advertising as well. You can use A/B testing to evaluate different promotions and ad types. For instance, it can help you see whether your customers value a percentage off more than a dollar amount (e.g. $20 off a purchase of $100).
7. Web Pages and Presence
Finally, consider using split testing in your web pages and web presence. On your website, A/B testing can show you where your customers prefer your menu bar, how many menu options they would like to see, which pages they find valuable, and what style of organization they like best – and that’s not all. Split testing can help you tweak your social media presence. Use A/B testing with different biographies, profile pictures, and cover photos to see what works for your company.
A/B Testing Best Practices
Whether you have questions about content, visual appeal, or how to word your CTA, A/B testing can help. However, the way you do it matters significantly. Follow these best practices to reach your customers more effectively and close sales more efficiently.
Pick One Variable
Start by selecting one variable. Do not, under any circumstances, attempt to test two or more variations at once. A/B testing answers only one question: Which is more effective: version A or version B? That’s it. If you try to tests more than one variable at a time, you won’t know which variable is driving the success or relative failure of the version you are testing.
For your first foray into split testing, start with something simple, like the color your use or the background image. From there, you can move onto many different variables, including the content itself, but for now, this will help illustrate the concept.
Establish a Control
Next, establish a control. This is usually your current site, page, post, or ad. You will apply the variation to this version.
Selecting a control is important because you cannot evaluate your changes effectively if you make too many changes at once or if those changes do not follow a linear course. Make all changes against this first version or control.
Two by Two
Further, do not try to run more than two versions at the same time. You want to test these variations against one another, so only use the control and a single variation. if you try to run several variations at once, you may not be able to tell which version is truly more effective.
Create Random Samples
To test your two versions, you will need a way to randomly select users. If you use a software to help you with A/B testing, it will usually (not always) do this for you. For instance, if you decide to run two ads on Instagram, you would make sure that you select the same demographics for each campaign. If the demographics you are testing are not randomly divided between the two version, your results will not be reliable.
Put both versions of your page, post, or ad out there, then sit back and collect the data. Give it some time. A week is good, two is better. You want to allow ample time to see whether a change is significant.
Finally, define what you mean by significant. Is the mark of a good ad that it got someone to click on the link? Or, do you want to measure how many of those conversions lead to sales? There are no wrong answers, but it is important to understand what you mean by success.
Why Should You Use A/B Tests
A/B testing helps you understand what works best. By testing two variations of a specific product to see which one gets you the desired outcome, you can isolate elements that are working for you. Split testing also helps you identify where you can make changes that could improve your conversion, sales, and other successes.
One of the biggest benefits you will enjoy from using split testing is that you will be able to make informed decisions. When you design something – anything – many of the decisions you make are arbitrary. You pick a font that you think looks good, a color scheme, or a format you predict your site’s visitors will like, but you probably did not test those variables – maybe they work, maybe they don’t, or maybe your content is compelling regardless.
However, what if you could know what your customers prefer?
With A/B testing, you can. It will help you determine the best font to use, the most effective color scheme, the best image, the right wording, where you should put the CTA button, how you should word your call-to-action, and so on.
In helping you find the right approach, split testing effectively helps you increase traffic to your site. Users find it more visually appealing or that the content is in a format that is more easily digestible.
Furthermore, you can continually tweak your designs and other efforts so that you compound the benefits of your efforts. Conversion rates can grow, then grow again with each new iteration of your split testing.
Putting it Together
A/B testing is effective, but it can be cumbersome. When it gets to be too much, keep in mind that the wisdom you ascertain from the exercise can be used again in countless ways. If you find that your followers like reading articles with bullet points, you can start including that feature in all your content. Also, remember that split testing takes a fair amount of organization to plan, execute, and oversee the testing of your variation pairs. Don’t hesitate to use a marketing data science provider like ironFocus to help streamline your efforts.