A/B testing is the ultimate experiment for user experience across your website and online marketing efforts. By splitting new visitors between two different versions of your website you can test specific elements, compare them and improve your conversion rate, based on real-life user data.
Not sure where to put that new email sign up form? No problem! You can run two versions and see which layout gets more people signing up – that’s the beauty of A/B testing.
Where Do I Start With A/B Testing?
Like all things in marketing, the first place to start with A/B testing is with some sturdy planning. First of all you need to know what you want to test and common elements include:
- Calls to action
- Headlines and content
- Website layout and design
- Product pages
- Pricing and promotions
Once you know what you plan to test, it’s time to make a list of variables. For the sake of example, let’s say you’re going to test the main call to action button on your homepage. Your list of variables will probably look something like:
- Location of the call to action
- The precise text it uses
- The button colour, design and surrounding space
This gives you three variables for your first A/B test and your goal is to see which combination gets the most conversions.
Choose Your Tools
One of the great things about A/B testing in 2015 is you have so many tools to help you through the process.Platforms like Optimizely mean you can make changes and run tests without touching a single line of code; which is great, because it’s best to start off simple.
But if you need something a little more advanced right away, then you might want to get in touch with our teamfor some advice.
Some Essential Dos And Don’ts
A simple test like our call to action example shouldn’t cause you any problems – but, just to give you a head start, here are some dos and don’ts for any A/B test worth knowing.
- Never test two variations independently – always run both elements at the same time, or your data becomes unreliable.
- When your test is over, don’t let your gut feeling or preferred option overrule test results. A/B testing throws up some surprise winners, but you don’t want to ignore these when profit is on the line.
- Don’t surprise regular visitors by sending them to modified versions of your website – especially when the changes may never become permanent. Only split new visitors between modified versions of your site.
- Don’t worry if your first split test doesn’t yield anything impressive – it takes multiple tests to gain real insights and improve conversions. So keep going!
- Know when to end your test. If your test is too short then your data isn’t trustworthy, while overrunning means the losing variation can cost you conversions.
- Show repeat visitors the same variation or you’ll only confuse them. Most A/B tools come with a mechanism to keep track of visitors and which variation they have seen.
- If you’re testing something that shows up on multiple pages or parts of your website, keep variations consistent everywhere.
- Always consider outside factors. Public holidays, the economy and even the news can have an effect on your industry or retail as a whole.
We’ll Be Back With More On A/B Testing Soon
Conversions are very important to us at Auburn and we’ll be back with more posts to help you improve the performance of your website with methods like A/B testing – so stay tuned.