Abstract

The research deals with the mathematical model of multivariate testing of the landing page of the website. The confidence intervals for the conversion rate difference of the landing page variations are investigated.

Highlights

  • The activity of each company is aimed at making profit

  • The existing of the conversion rate difference outside the limits of the confidence interval when the number of visitors is reached 14890 in each group means the rejection of the hypothesis H0ij

  • The existing of the conversion rate difference outside the limits of the confidence interval when the number of visitors is reached 14835 in each group means the rejection of the hypothesis H0ij

Read more

Summary

Introduction

The activity of each company is aimed at making profit. At a time when much of the target audience gets information about products and services on-line and making purchases on-line, it is important to have an effective website for every company. All possible combinations of landing page elements are simultaneously tested with multivariate testing It helps to evaluate the impact of each element and their interaction on the conversion rate. The landing page variation (optimal combination of elements), which won the greatest conversion rate is chosen according to the test results. It should be noted, that multivariate testing is expensive and takes time. The Bayes’ theorem is used to find a posterior probability distribution of parameters that is using in Bayesian point estimates calculation and in building confidence intervals for unknown parameters In some cases it is available the information about a priori distribution with the accuracy to unknown hiperparameters that can be evaluated simultaneously with the assessment of unknown parameters. Confidence intervals limits for the unknown parameters are calculated solely on the sample and they do not depend on unknown parameters, and the intervals themselves contain unknown parameters with a set probability

Goal setting
Terminology
Multiple hypotheses testing under the multivariate testing
Implementation of multivariate testing
Conclusions
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call