A HADI cycle is a hypothesis testing tool. It helps understand how business is affected by changes. This tool allows you to track the impact of updates on business performance and avoid situations where it's unclear why numbers have changed.
HADI cycles are used in digital marketing to measure performance of ad campaigns, website, and other communication channels.
A HADI cycle consists of 4 steps:
1. Hypothesis
First, a hypothesis to be confirmed or disproved is made. Come up with a hypothesis that is based on customer interviews, stats, or data from previous experiments, and not on your gut feeling.
The hypothesis must be specific and measurable. Choose one metric that should be affected by the change. E.g.: If we change the heading on the home page, the conversion rate increases by 20%.
2. Action
Now you want to implement the hypothesis, so you actually modify that heading. It’s crucial to stick to the initial experiment idea.
You can test several hypotheses in one cycle, as long as they affect different metrics and don't involve a single segment of users.
3. Data
It’s time to track how the selected metric has changed over the chosen period of time. To do this, conduct an A/B test.
Tip: Do the A/A-test first to ensure that the chosen tool is good at randomizing users.
4. Insight
After collecting the data you need to analyze it and see if the hypothesis is confirmed. If so, keep the change or scale it up. If the hypothesis is not confirmed, i.e. metrics have worsened or performance remains the same, roll back the update and test another hypothesis.
💡 Gather and analyze the data accumulated over the course of the experiment. Think of what you have learned that was not obvious before. Use your learnings to create future hypotheses.
HADI cycles help businesses quickly test ideas and eliminate ineffective ones. To test multiple hypotheses, start with the ones that are easier to implement and are expected to have the greatest positive impact.
HADI cycles are used in digital marketing to measure performance of ad campaigns, website, and other communication channels.
A HADI cycle consists of 4 steps:
1. Hypothesis
First, a hypothesis to be confirmed or disproved is made. Come up with a hypothesis that is based on customer interviews, stats, or data from previous experiments, and not on your gut feeling.
The hypothesis must be specific and measurable. Choose one metric that should be affected by the change. E.g.: If we change the heading on the home page, the conversion rate increases by 20%.
2. Action
Now you want to implement the hypothesis, so you actually modify that heading. It’s crucial to stick to the initial experiment idea.
You can test several hypotheses in one cycle, as long as they affect different metrics and don't involve a single segment of users.
3. Data
It’s time to track how the selected metric has changed over the chosen period of time. To do this, conduct an A/B test.
Tip: Do the A/A-test first to ensure that the chosen tool is good at randomizing users.
4. Insight
After collecting the data you need to analyze it and see if the hypothesis is confirmed. If so, keep the change or scale it up. If the hypothesis is not confirmed, i.e. metrics have worsened or performance remains the same, roll back the update and test another hypothesis.
💡 Gather and analyze the data accumulated over the course of the experiment. Think of what you have learned that was not obvious before. Use your learnings to create future hypotheses.
HADI cycles help businesses quickly test ideas and eliminate ineffective ones. To test multiple hypotheses, start with the ones that are easier to implement and are expected to have the greatest positive impact.
