Using A/B Split Testing To Improve Website Results
Creating a great website consistently comes down to incremental improvement. Anything that is designed to be used by people needs to be tested and tweaked to get it just right. A/B split testing is one of the best methods to accurately measure if a website change is having a positive effect.
Put simply, A/B split testing on websites involves creating a copy of the page you want to improve, then making a tweak to your new page. When visitors come to your website, they'll be randomly presented with one of the versions. This way you can track the data and user behaviour that comes from each of the variations, and use it to figure out which one is more effective.
There are a range of tools available to help you do this and track the results. Popular choices include Optimizely, Visual Website Optimizer and Google’s free Content Experiments which is built into Google Analytics. Most websites already use Google Analytics for overall data tracking and management, so this is usually the easiest method of getting started with split testing.
Found under the Content section of the left hand side menu, Experiments uses a guided creation wizard to get your experiment set up. All you need to do is enter the URLs of the variation pages that you want to test, name and describe your experiment. It will automatically give you code to add to your main control page, and then will automatically compare the results of the variation pages that you specify. This traffic will then be tracked like regular Analytics data, and most importantly you can set up a goal in order to see which changes are making the most conversions.
Some ideas for split testing include:
Best practice in split testing technique is to make a variation on only one area at a time. This way you'll understand exactly which changes have made the difference. Don't try to contrast apples and oranges - if there are multiple areas on your website that you want to experiment with, then keep them for separate tests to ensure your results don’t become muddled and confused.
Usually we just follow our gut feel on what looks best, but it’s not always the case. There can often be surprising results. Split testing is an excellent way to narrow down your design choices and get actual practical feedback about which decisions are actually the correct ones.