This paper develops a flexible methodology to analyze the effectiveness of different variables on
various dependent variables which all are times series and especially shows how to use a time
series regression on one of the most important and primary index (page views per visit) on
Google analytic and in conjunction it shows how to use the most suitable data to gain a more
accurate result.
Search engine visitors have a variety of impact on page views which cannot be described by
single regression. On one hand referral visitors are well-fitted on linear regression with low
impact. On the other hand, direct visitors made a huge impact on page views. The higher
connection speed does not simply imply higher impact on page views and the content of web
page and the territory of visitors can help connection speed to describe user behavior.
Returning visitors have some similarities with direct visitors.