What is A/B Testing?
An A/B test is an experiment where you test one or more variant changes on a website. Variant A is the original and variant B is the modified version. Once you establish what you want to change, then you create a hypothesis of what you predict will happen when you make the change. It’s important not to change too many things at once or your data wont be able to pinpoint what is causing the results to change. After you create your hypothesis, then you are able to test. To run the A/B test, simply go to google optimize and create the test. The help page recommends keep the experiment running until “Two weeks have passed, to account for cyclical variations in web traffic during the week. Or, at least one variant has a 95 percent probability to beat baseline.”
Sample Hypothesis:
Looking at increasing scroll depth, one thing that I would like to test is changing the location of the YouTube video on the page will increase the amount of 100% page scrolls. With that said, My sample hypothesis is that if we move the YouTube video to the bottom of the page then the percentage of users who scroll all the way down the page will increase by 15%. The next step will be for our team to run this test through google optimization and see if the variant change effects the page scrolling depth.
https://support.google.com/optimize/answer/6211930#create-a-hypothesis