In general, a hypothesis is a proposed concept or idea that serves as a starting point for experimentation. In A/B testing, a hypothesis is a specific change that can be made on your site to improve a tangible metric. It should be outlined as an If-Then-Because statement. If you make X change, then it will cause Y outcome because of Z reason. A hypothesis can be proven correct or incorrect through testing.
How to Construct a Hypothesis
Questions to think about:
- What are current visitor pain points?
- How could you help customers find what they are looking for?
- What might bring more visitors to a particular page?
A hypothesis should be limited to a singular change, but there could be multiple variations to that change. If the hypothesis is too broad, you will not be able to say with accuracy that a specific change caused the outcome.
Find some example hypotheses here.
Where to find insights
Insights can come from lots of places. You may have an informed idea based on the results from a former experiment. You may have spoken to your customers directly in interviews or focus groups. Or you may utilize analytics tools to capture insights that can then translate to an A/B test hypothesis.
Popular tools for gathering these insights include GA4, heatmapping tools, customer survey tools, and more. Learn how to integrate Compose with GA4.