Split-testing is an incredibly powerful, important, and often under-utilized tool. Though once limited to big companies like Amazon, split-testing has been steadily increasing in popularity among a variety of businesses, including SMBs. In fact, split-testing can be incredibly simple if you have the right tools, and you can use it to increase the effectiveness of your online presence.
Yesterday, we held a webinar with MixRank on how you can get creative with split testing for more effective marketing. Here’s what you may have missed:
What is split-testing?
Split-testing is an experimental approach to web design that allows you to test hunches about what will perform better. This is done by comparing a control of a public property to a test version. If this is totally new to you, heres a simple example that should help: You want to find out if removing a form field from your landing page will increase conversions. Your instinct tells you it will, but you want to be sure. Instead of relying on your hunches, create a second version of your landing page with that one field removed. Send half of your users to Version A (with form field), the control, and half to Version B (without the form field), the variant.
What makes for a successful A/B test?
Fist, set clear, measurable goals. In order to be successful, you have to know what success looks like. This will most likely vary from test to test. Select quantifiable metrics to measure your winner. For example, will conversions, clicks, or impressions serve as your metric? Whatever youre testing, you always want to make sure you have one control, version A, and one variant, or your Version B. Its important to only test one variation at a time so you know which variation led to an increase or decrease in conversion rates. Always randomly divide your users into buckets to make the test truly experimental. The people in your test must be divided into different buckets at random. Do what you can to ensure that there are no other factors that could be influencing differences in performance.
When and how can I use split-testing?
While you can test for virtually any aspect of your web presence, most marketers test emails, landing pages, and advertising. You can test email subject lines, calls-to-action, image placement, length of text, copy, buttons, and more. If you want to know which email will have the best open rate, offer two drastically different subject lines, and then measure your results. Are you wondering which call-to-action has the best click-through rate on your landing page? Do you already know which call-to-action is best and now you want to know which color the button should be? These are all perfect examples of when and how to use split-testing.
Best practices of split-testing?
- Only test for one variable at a time. If you have two variables, you’ll be unable to attribute an increase or decrease in conversions.
- Go big. Don’t be afraid to make drastic changes. A completely new value proposition, for example, can make a big difference.
- Don’t quit. As long as there is room for improvement, there is always room for testing.
- Set clear and measurable goals. It’s all about how you define success.
- Never end a test prematurely. You want to make sure your test is statistically significant before cutting it off.
Here at ReTargeter, we always run A/B tests and recommend that our clients do the same. Most recently, we tested the ad headlines, “Retargeting Made Simple” vs. “Retargeting Made Easy.” Our value proposition for this campaign was the simplicity and ease of use of our product. So, we created two headlines to communicate that notion. The next time you’re running a marketing campaign, don’t rely on what you think will work, and instead run a test for measurable and actionable results.