How to Use A/B Testing to Increase Conversions

A/B testing compares two slightly different variations of a landing page, online advertisement, email, etc against each other to determine which one performs better. Also called split testing, A/B testing is an excellent way to validate new design changes and improve conversion rates on landing pages. The variations on each can be as simple as a change in the headline, a different style or colour of the button, a different call to action or image placement.

Once implemented, half of your traffic will be shown the original version of the page (known as the control) and half will be shown the modified version of the page (the variation). As visitors are served either the control or variation, the performance is measured through data and statistics using tools such as Google Analytics, WordPress plugins, or paid software such as Clickfunnels or Leadpages — so there’s no more guesswork required! You will have actual data to shift the conversation from “We think this design will work” to “We know this design is working”. The end result is ultimately a better conversion rate. Are you excited yet? You should be, so let’s dive into the A/B testing process.

Before running an A/B test you have to understand which page(s) need help. Dig into your analytics and look for pages with low conversion rates or high drop-off rates. Remember, opportunity doesn’t mean fixing your worst performing pages. It means testing pages that have the biggest impact on your bottom line. So be sure to look for the pages that are tied to your sales opportunities.

Now that the page or pages with improvement opportunities are staring you in the face, it’s time to start your A/B test. There are six steps involved and please keep in mind that this is an iterative process. You may not get it right the first time, but if you continue to follow through with these steps, you will see results.

 

1. DETERMINE GOAL FOR PAGE

Now, identify your goal for that page. Do you want someone to click a button? Should they fill out a form or opt-in? Do you want them to download a sales sheet, guide or offer? This goal will become the key metric that you set to determine whether or not the variation is more successful than the original version.

2. DEVELOP HYPOTHESIS

Once you’ve identified your goal, develop a list of ideas to make the page variation better than the control page. It’s okay to keep this list simple. Remember: even small changes, like switching the colour of a green button to a red button, can impact the conversions. In order to keep your A/B test simple and ensure that your results are statistically significant I highly recommend that you only test a control against one variation and only make one change per test. If your list is long then expect to run several experiments.

3. TEST HYPOTHESIS

In order to test your hypothesis I recommend addressing one item at a time. If you try to experiment with too many changes all at once you will have a hard time determining which change actually made a difference. This might be changing the colour of a button, swapping the order of elements on the page, hiding navigation elements, or something entirely custom. Not a web designer? No problem: there are many landing page tools out there that allow you to build landing pages using drag and drop editors – UnBounce, LeadPages and ClickFunnels are by far my favourite.

Once you’ve made the changes, set up your preferred A/B testing software to display your control and variation pages equally but randomly to your website visitors. Using this method 50% of your visitors will see the control page and 50% will see the variation page. Their interaction with each page is measured, counted, and compared to determine how each performs.

4. COLLECT AND ANALYZE DATA

The last step is to collect data and analyze the results with a focus on the data that directly relates to the goals of your page. One question I often get asked is “How long should I run my A/B test for? The answer is – it depends. The duration of your test is less about time and more about the statistics you need to make an informed decision. In a technical article about How A/B Testing Math Works Alex Castrounis states:

The goal is to make sure you collect enough data points to confidently make predictions or changes based on the results. While the math behind determining the appropriate number of samples required for significance is a bit technical, there are many calculators and software applications available to help. For example, evanmiller.org has a free tool you can start using right now.

The point being if only 100 people visit the pages you are testing and neither get a conversion, you have zero data to make a conclusion. So, you need a reasonable sample size (in this case, traffic to that page and some conversions) to make conclusions. Also, you need to consider how many combinations you are testing. The more variables, the longer you will need to test.

5. MAKE CONCLUSIONS

Based on the data collected, determine which page variation performed better. This conclusion can be as simple as: “After 500 visitors to each page it was observed that the green button outperformed the red button. We believe this is because the green button has better contrast against the other colours on the page than the red button.” The beauty is that these conclusions are no longer just based on theory — they are based on actual hard numbers thanks to your A/B test…so give yourself a pat on the back.

6. RUN NEXT EXPERIMENT

This is a never ending, cyclical process of making a hypothesis, testing, analyzing data deriving conclusions and then moving on to the next item that you believe will help improve the page. For example, you can now try running yet another experiment, this time changing another part of the page, i.e. the call to action, the headline, the form placement, and see if that helps the conversions increase even more. Eventually you will reach the holy grail of a high converting landing page.

Now that you understand the process, you are probably wondering when it’s worth implementing an A/B test. We recommend A/B testing whenever you aren’t achieving your desired results from a landing page, online advertisement or email. As you can see by the process, it’s not complicated to do and you can see the results within a short time frame.

If you have a page you’d like to test, contact us and we will discuss opportunities for A/B testing that could help you generate more conversions.