Is A/B testing right for you? Let’s say you have a new content marketing campaign where you have created a fantastic landing page where a visitor has arrived through a link in your blog, a banner ad, an email or your website. How can you tell if this new landing page is working as effectively as it could be? A-B testing, or testing two slightly different versions, can give you the answer. What kinds of things can be tested that may make a difference – even a little one? Here are a few quick tips that should help:
Remember these 10 Tips when A/B Testing
One test at a time
1. Make sure you only conduct one test at a time. If you test an email campaign that directs to a landing page at the same time that you’re A/B testing that same page your results will be skewed since you won’t know which change was making the difference.
One variable at a time
2. To properly evaluate the effectiveness of an element on your page (e.g. Click Here button), then isolate this as the only thing that will be tested. Attempting to test more than one element will again result in an inaccurate account.
Test minor changes
3. Often times the biggest differences are created by doing the smallest of things, such as changing the color of the call-to-action button. Think about the small things first and see what happens.
Test the whole design
4. While testing the color of a call-to-action button will yield results, you can also think about presenting two different page designs. A/B testing two different layout options can have a positive effect on the bottom line – so it is worth considering.
Measure as much as you can down the sales funnel
5. As you create your A/B testing, keep in mind what factors are causing greater click through rates. Take note of the paths people took where you made a sale. What were they looking for?
Set up control and treatment
6. Always make sure you keep your original version as a control so you can return to “square one” if your tests do not produce positive results.
Decide in advance what you want to test
7. There is more to A/B testing than just changing the color or size of buttons. Look at the fonts and point size as well as background colors. And test these changes and edits with different target audiences using different methods (email, blogs, banner ads) to get them to visit your landing page.
Split your sample group randomly
8. A/B testing should be down with two or more groups that are roughly equal. For instance, if you are sending out emails, make sure the lists have approximately the same number of recipients. And you must split these lists randomly or your results will not be conclusive.
Test at the same time
9. Timing plays a large role in determining the effectiveness of A/B testing. Try to pick a day of the week and a time that is consistent.
Decide on significance before testing
HOW WE TRIPLED ECOMMERCE REVENUE IN 18 MONTHS
By capitalizing on conversion-centric email marketing, social media, online giveaways, and abandoned cart emails, we significantly increased brand loyalty, awareness, and most importantly—eCommerce sales.
10. Before commencing your test be sure to set the statistical significance goal; 97-99% statistical significance is optimal as a good percentage to aim for.
Remember all of these tips will help make your A/B testing much more valuable and will hopefully yield profitable results.
Snackchat! Taco Bell tweeted, “We’re on Snapchat” last night, urging their Twitter followers to add them for a “secret announcement” sometime today.