Why & How to Split Test a Website (A/B Testing)

Part of clever marketing is understanding what works best. But how do you really know what type of website and content will convert? The experience will get you so far, but to really deliver results you need to understand about testing. It’s an old theory, but trial and error is sometimes the best source of marketing data you can have.

What is Split Testing?

Split testing is the process of testing two similar pieces of content with one variable to see which performs best. For websites and landing pages, this means creating two versions with a slight difference and testing them simultaneously to determine which converts more or performs best.

You may also have heard it called A/B testing because you’re looking at Sample A and Sample B to test which works better. Depending on your goals, the variable you choose to test could be anything. It may be testing two colour schemes or even font types. It could be different placement or design of Call to Action buttons. There’s no limit to the variables you can test, but it’s a matter of knowing what you want to achieve.

Why is Split Testing Important?

Split testing is important because it gives marketers a chance to try different concepts to deliver better results. If you’re running Google Ads or social media ad campaigns, it’s good practice to run several at the same time to see which one generates more leads. This gives valuable insight into where your marketing dollar can be best spent in future.

Split testing of websites is simply the next logical step but focuses on more than just lead generation. Businesses may want to find out which type of landing page converts better or which page design reduces bounce rates. eCommerce stores may focus on checkout page design to reduce the number of abandoned shopping carts. Ultimately, it’s all about gathering data to make your future marketing spend more effective.

How to Split Test a Website

Now you know why split testing is so important, here are the easy steps to follow in order to conduct your own.

1. Understand Your Goals

What do you want from your website? More conversions, increased traffic or something else? Once your testing is done, you can certainly analyse a range of different metrics, but it’s ideal to have one clear goal you’re trying to achieve. This better informs the variables you choose.

2. Choose a Single Variable to Test

Firstly, it’s important to only test one variable at a time. This is so you get a clear picture of what works best. If you have more than one variable, it’s impossible to determine which factor has caused the best results. So, settle one variable that aligns with your goals.

3. Create Version ‘A’ and Version ‘B’

With your variables chosen, now it’s time to create version A and B of your website or landing page. You can create two completely new ones, or you may choose to use your existing page as ‘A’, and duplicate it with your variable as version ‘B’.

4. Set Up a Split Testing Tool

If you’re operating a CRM, it’s worth checking what internal tools you may already have for split testing. Many have built-in testing tools which can deliver, test and report on different pages. If not, you can access a tool such as Google Analytics Experiments.

These tools are also important for ensuring the integrity of testing. For accurate results, both versions of the page need to be seen by roughly the same number of users. A/B testing tools deliver your page to a random sample of users.

5. Test Both Variables at the Same Time

Again, we can’t stress enough that slit testing must occur simultaneously. This keeps the integrity of your results because it captures a certain period of time where random users are viewing both versions of your page.

6. Measure Your Results and Take Action

Finally, it’s time to check your results. You already had your primary objective in mind, so this is the first result you check. Which version performed better for your goal? Don’t disregard other metrics though, because you need balance in your page’s performance. If version ‘A’ achieved your goal but suffered badly on other metrics, you can always resume split testing with another variable.

Once you’ve done your split testing analysis, you can use that information to better inform future marketing efforts. More leads, more conversions, more success, and most importantly, more cost-effective.

Need a hand with building and split testing a website that will convert? Contact Cemoh now.