A/B and MVT testing- pros and cons.
Split testing is an excellent tool that allows increasing the site conversion and its profit. The primary task in the organization of such experiments is the identification of significant factors (for example, subtitle, pictures, etc.), and their assessment of their impact on the conversion.
A large part of the literature, dedicated to this problem, places emphasis on the statistical analysis, which scare many people away. Thomas Høgenhaven tried to create more affordable illustrated guide describing how to properly organize the split testing and assess the results obtained.
A/B or MVT
First of all, you need to define the general testing system. To assess the impact of one or two factors, people usually use A/B tests. If there are more factors, it would be better to use the multivariate testing (MVT). We should note that using the series of A/B tests, you can also check two or more factors, but in this case, the information on their interaction will be missed.
Variants of MVT: full-size and partial
If you have opted for the multivariate testing, you have to choose the specific option. The most common variants are: full-size and partial.
Checking the test condition by means of A/A test
Most of the aspiring automation testing companies are familiar with the term “A/B test”. But only of them know about A/A test. It is used for checking the test conditions. A/A test should be performed before starting A/B or MVT tests. A/A test must show that the visitors have been correctly divided into groups, and no other factors (except for the tested ones), will not affect the result.