Adding a hypothesis to an experiment in Sitecore Personalize
The purpose of adding a hypothesis is to document what the anticipated outcome of the experiment is, while enabling others to view your expected results. Each hypothesis is an idea to be tested.
The first step of any experiment or A/B test is to formulate a hypothesis that you want to test. We recommend adding a hypothesis before adding goals.
Although optional, declaring a hypothesis is the basis for a good experiment because it ensures that the marketer can summarise the intended outcome, and measure against it. For example, a basic hypothesis might be that “Variant B of the home page performs ‘better’ than the current one (which is the control)”. An even better hypothesis is “Variant B of the page will lead to at least 20% more conversions than the control”.
A hypothesis written this way can result in one of two outcomes:
-
The test proves there is no difference between Variant B and the control, which results in a null hypothesis.
-
The test is successful, where Variant B performs better against the control, which results in the alternative hypothesis. The alternative hypothesis is what you hope your Experiment helps prove.
Apply the following when declaring a hypothesis:
-
Avoid ambiguity in terms of cause and effect
-
Test only one change at a time
-
Estimate the desired effect