Here is how you can identify the average number of users accessing the system during peak hour from Google analytics statistics.

Pick up the peak traffic day statistics & capture the below metrics.

  • Average Session Duration = 7minutes 54 seconds = 474 seconds
  • Average Pages visited by User = 7.89 Pages / session
  • Hence, Average time per Page can be calculated as 474 / 7.89 = 60 seconds. (This includes both response time & think time)

Pick up the peak traffic hour (July 14th - 13th hour) Statistics :

  • Total User sessions during peak hour= 1312 Users
  • Total PageViews during peak hour = 9359 Pageviews / hour = 9359/3600

= 2.59 Pageviews/second

By Little’s law, N = X * R

Note : Take enough care in the unit of measurements used for metrics – N, X & R

  • Option 1 : Apply Little’s law (using average time per page & individual page throughput) :

Average number of Users at any time during peak hour = 2.59 * 60 seconds = 155.4 users

  • Option 2 : Here is another way to apply Little’s law (using average session duration of a user & individual page throughput):

Average number of PageViews at any time during peak hour = 2.59 * 474 = 1228 PageViews

Hence, average number of users at anytime during peak hour = 1228 / 7.89 = 156 users

This means on an average, there were 156 active user sessions during peak traffic hour. Hence, you need 156 virtual user licenses to simulate this load (for using realistic think times).

Note : The above calculations can be performed for load test conducted in LoadRunner / JMeter tool.

For more detailed explanation & practical case studies,you can refer to our course “Workload Modelling Essentials for Performance Engineers” . This course explains workload modelling & performance test validations using Little's law with practical case studies.