How to do A/B testing

There are numerous shared reports of A/B testing online, each showing varying results from the process. What they all have in common is that they manage to do in a few hours or days what well-funded marketing teams couldn’t do in years. Users report doubling conversion rates by testing different home pages. Variations in images, titles, calls to action, button colours and placement all make a difference to how visitors react. Tools are now available that require low technical ability to implement them and start testing immediately.

If carrying out a test, be sure to measure conversions and sales, as increased click-throughs alone are not always indicative of a complete conversion.

Videos, smiling human faces (it’s true!) and certain power words (free, now, instant etc.) are all said to convert more visitors to clients than non-optimised text. Take a look a few alternatives such as Optimizely, Visual Website Optimizer (featuring a great blog with many case studies), Unbounce or to find out more on the subject.

There can be statistical grey areas in A/B testing – a sample size that does not include enough visitors could give insignificant results, as could a larger sample size if the tests vary over multiple differences. If you would like to delve deeper here, you can use a significance calculator.

Thanks for reading. I do translation from French and Swedish to English, so if that's useful to you, feel free to connect and message me on LinkedIn or Twitter.