Multivariate Analysis Methods
Register for the Ryte Newsletter
Get the latest SEO and website quality news! Exclusive content and Ryte news delivered to your inbox, every month.
Multivariate analysis methods are used in the evaluation and collection of statistical data to clarify and explain relationships between different variables that are associated with this data.
Multivariate tests are always used when more than three variables are involved and the context of their content is unclear. The goal is to both detect a structure, and to check the data for structures.
Multivariate analysis methods can be used to systematically increase the usability of websites. While A/B tests always isolate only one web page, multivariate methods show the relationships and interactions of several elements within a web page. The significance depends on which and how many elements of the website are used. All elements of the website that enable the user to interact with the website via the user interface are generally considered variables. This includes in particular those that have an impact on the conversion rate.
Originally, multivariate test and analysis methods were used in statistics to uncover causal relationships. Since manual calculations are very complex, the methods only became practicable in other fields of application with the development of corresponding hardware and software. Multivariate analysis methods are used in a variety of areas:
- Linguistics, Natural Sciences and Humanities
- Economics, insurance and financial services
- Data mining, big data and relational databases
Multivariate analyses are usually carried out using software in order to deal with the huge amounts of data and to monitor the changed variables in practical applications such as usability tests. However, multivariate tests can also make a significant contribution to improved user-friendliness on a smaller scale.
Types of multivariate analysis methods
Multivariate methods can be subdivided according to different aspects. First of all, they are differentiated according to whether the aim is to discover a structure within the combination of data, or whether the data is to be checked with a certain structure. a structure The structure-determining methods include:
- Factor analysis: Reduces the structure to relevant data and individual variables. Factor studies focus on different variables, so they are further subdivided into main component analysis and correspondence analysis. For example: Which website elements have the greatest influence on purchasing behavior?
- Cluster analysis: Observations are graphically assigned to individual variable groups and classified on the basis of these. The results are clusters and segments, such as the number of buyers of a particular product, who are between 35 and 47 years old and have a high income.
Structural review procedures include, among others, the:
- Regression Analysis: Investigates the influence of two types of variables on each other. Dependent and nondependent variables are spoken of. The former are so-called explanatory variables, while the latter are explanatory variables. The first describes the actual state on the basis of data, the second explains this data by means of dependency relationships between the two variables. In practice, several changes of web page elements correspond to independent variables, while the effects on the conversion rate would be the dependent variable.
- Variance analysis: Determines the influence of several or individual variables on groups by calculating statistical averages. Here you can compare variables within a group as well as different groups, depending on where deviations are to be assumed. For example: Which groups most often click on the' Buy Now' button in your shopping cart?
- Discriminant analysis: Used in the context of variance analysis to differentiate between groups that can be described by similar or identical characteristics. For example, by which variables do different groups of buyers differ?
A multivariate test of a web page can be presented in the following simplified way. Elements such as headlines, teasers, images, but also buttons, icons or background colors have different effects on user behavior. Different variants of elements are tested. The test would initially identify these elements and show different users differently designed elements. The aim would be to obtain data on the effects of the changes in terms of conversion rate or other factors such as retention time, bounce rate or scrolling behavior compared to other sets of elements.
Significance for usability
As a quantitative method, multivariate analysis is one of the most effective methods of testing usability. At the same time, it is very complex and sometimes cost-intensive. Software can be used to help, but the tests as such are considerably more complex than A/B tests in terms of study design. The decisive advantage lies in the number of variables that can be considered and their weighting as a measure of the significance of certain variables.
Even four different versions of an article's headline can result in completely different click rates. The same applies to the design of buttons or the background color of the order form. In individual cases, it is therefore worth considering from a multivariate perspective also financially, especially for commercially oriented websites, such as online shops or websites, which are to be amortized through advertising.
- Multivariate Testing in Action: Five Simple Steps to Increase Conversion Rates smashingmagazine.com. Accessed on 04/10/2014