If you’re not A/B testing yet you should be, because you’re leaving money on the table. How do you make good business decisions–do we take guesses and settle for good enough? Nope, we get the data and A/B testing is how we do it.
What Is An A/B Test?
A/B Testing (also called split-testing) is the process of comparing two versions of a web-page, advertisement, email, or any other marketing asset. A/B Testing is when you test just 1 change, such as a button color or text size.
Here are some elements that are commonly tested in an A/B test:
- Button shape, color and placement
- CTA (call to action) text
- Copywriting
- Images/video used
- Body content
- Social proof
After you’ve prepared your two versions of your page (or other asset), present both versions to half of your visitors. The performance of each should be judged on metrics such as conversion rate, time on page, or button clicks depending on the type of content you’re testing.
Whichever version of your marketing asset performs better is the winner, and you can then test another aspect of it. For example, if you initially tested the CTA button color and the data shows orange performs best you can make a new A/B test to see whether medium-sized headings or large-sized headings convert better and make the CTA button orange on both.
When you want to test multiple things at once you have to create a version for each combination of the changed elements and it’s called multi-variate testing. You can think of multi-variate testing as A/B testing’s big brother, but for the purpose of explaining the benefits of A/B testing we’ll keep things simple. Wingsend does support both A/B testing and multi-variate.
The Commonality of A/B Testing
Okay so you’ve heard of A/B before and you’re not entirely convinced it’s for you or that it’s necessary at all. Well A/B testing is certainly a popular practice in both small business and large corporations–so if you think google knows what they’re doing it may be of value to your business. They did their first A/B test in the year 2000 (to determine the optimal number of results to display).
Here are some interesting facts:
- Improving your User Experience Design as a result of your A/B testing can increase your conversion rate by 400% (Entrepreneur, 2018)
- 71 percent of companies run 2+ A/B tests a month (Invesp)
- Obama’s campaign raised an estimated additional $75 million as a result of their A/B testing (Wired, 2012)
- Microsoft Bing’s revenue was increased by 12% from a single change in response to an A/B test (Harvard Business Review, 2017)
- Amazon.com, Google, & Booking.com conduct 10,000 A/B tests each per year (Harvard Business Review, 2017)
When You Should A/B Test?
Whenever you’re making a change to the user experience you should be measuring the impact of your changes. After all, you’re likely redesigning in an effort to increase your performance of that web-page, advertisement or other digital asset. Here are a few situations where you may want to A/B test:
- Website Redesign
- When you change a feature of your website, plugin or service
- Pricing change
- Conversion rate is too low
- You just want to raise revenue
Though there are a few caveats, it’s generally in your interest to make data-driven decisions rather than just going with gut feelings. For instance, you never want to redesign your website just because you think it’s too old and not hip enough. Just ask Digg whose unnecessary and constant redesigns annoyed its user-base and lost them traction in a war with Reddit. If you’re unfamiliar with Digg, you can guess who won that war.
Important Caveats
On average, an A/B test will require at least 25,000 visitors to reach a statistically significant sample size. What does this mean for smaller businesses hoping to test their landing pages, emails, and ads? It means that if you’re using paid traffic to drive conversions, you can likely hit that number of visitors (even if you have to stretch your test over a few weeks or a month). If you’re using organic traffic and not able to get that many visitors in a reasonable time-frame then don’t A/B test; focus on increasing your traffic first!
After all, A/B testing is all about optimizing the experience for existent users and increasing conversions–but no amount of optimization would be helpful if you don’t have enough traffic to begin with. Split Testing (A/B) is great but it’s one of many tools you need to utilize.
Conclusion
When your pages are converting better, your revenue goes up–it’s this universal truth that’s responsible for the popularity of A/B testing. Wingsend provides A/B testing for landing pages and emails deployed from our platform, and we have a low one-time fee that will save you thousands in monthly fees from MailChimp, Hubspot, and other services you’d need to have all of our features.
A/B testing is an essential tool to have in your toolbelt as your company grows, and even the simplest change can have unexpectedly powerful results so test everything. Just remember to follow the data and let your tests run long enough to reach a statistically significant sample size. Here’s a statistical significance calculator for your convenience! Contact us if you have any questions about Wingsend.