A/B testing takes out the gambling when deciding to make changes to your website. By comparing two or more variants we discover which version works best and produces the most conversions. In this way, your decisions are based on the data you have collected during testing and not on internal opinions. By measuring the impact that changes have on your KPIs such as subscriptions, purchases, downloads or other objects; you are sure that your website will evolve positively.
A/B testing is not a one shot, but a constant process of determining test opportunities, hypothesis formulation, variant definition/design, testing, evaluation and adaptation. The focus must remain on the general objectives of the website, even if only small parts of a funnel or process are tested. A well-defined test framework is therefore essential to get real value from A/B testing. The realization of the framework consists out of three steps:
1. Defining conversion optimization goals
2. Prioritize optimization goals
3. Identify the first test capabilities
The set-up and execution of the tests
An A/B test is only useful if you know what you want to test, and why you want to test it. Therefore, we always formulate a hypothesis before each test. This hypothesis is based on previous experience or on a theoretical framework.
On the basis of the hypothesis the variables are created and the test is configured.
Reporting test results
After the test Humix provides a report of the results. On this basis, the hypothesis can be accepted or rejected, and our consultant will make recommendations to modify the website, or to perform a follow-up test.
Practising the best variant
Once the winning variant is chosen, it will be developed and the adjustments can be made on the website.