This article describes how time spent browsing is calculated. This is referred to as Unit: Time spent in the Smoothwall Filter and Firewall. Custom reporting sections offer a reporting Option of Unit. This allows you analyze activity within domains by number of Hits, amount of Bandwidth used, and Time spent browsing that domain.
The Time spent option shows how much time was spent browsing a given site.
Report data is grouped into five minute windows. Within each time-window, hits are further grouped according to domain, for example, 20 hits to Facebook, and 10 hits to Twitter.
It should be noted that each five minute window can account for more than five minutes of real time. Clearly, a user cannot browse for more than five minutes in a five minute window. However, two users browsing to the same domain at the same time, may produce results of more than five minutes.
Let's take the example above of 20 hits to Facebook and 10 hits to Twitter, and two users, Dan and Tim. Within the five minute window, we record 15 hits to Facebook and 5 hits to Twitter from Dan, and 5 hits apiece from Tim.
This activity is calculated into a ratio of hits for each user:
site_hits / total_hits * time_window_for_each_domain_or_user_combination
Therefore:
- Dan (Facebook) :
(15 hits / 20 total hits) * 5 minutes = 3 minutes 45 seconds - Dan (Twitter) :
(5 hits / 20 total hits) * 5 minutes = 1 minute 15 seconds
- Tim (Facebook) :
(5 hits / 10 total hits) * 5 minutes = 2.5 minutes - Tim (Twitter) :
(5 hits / 10 total hits) * 5 minutes = 2.5 minutes
Combined together gives a total of:
- Facebook :
3 minutes 45 seconds + 2.5 minutes = 6 minutes 15 seconds - Twitter :
1 minute 15 seconds + 2.5 minutes = 3 minutes 45 seconds
within a single five minute window. This is added to all other 5 minute windows in the report time period to produce a grand total for the requested user, domain, or user per domain depending on the selected report options.
However, if a user has only one hit to a single website during that five minute window, it is assumed the user spent the whole five minute window on the one website. No hits to any website at all during a window assumes the user is not browsing.
The reporting engine also ignores background activities associated with websites, such as the updating of other browser tabs that are open, updating of image or advert carousels, and so on.
This method allows for time to be considered in the same way as hits and bandwidth, making reporting more flexible.