Anyone who runs a web application with serious intent is always interested in two things: How is my web application being utilized? And what is the experience for those users? The first question is answered by looking at data that tracks a user’s journey through the web application, while the second question is answered by looking at data that tracks page download speeds and application availability. Generating these two data sets requires two completely different tools – an analytics tool for tracking web application usage, and a performance monitoring tool for tracking web application performance.
However, there is also a hidden question: How do my users impact my performance, and vice versa? Combining the two data sets provides the answer! Here is a sample graph that combines performance data (as a line graph) with analytics data (as a bar graph):
The combination of these two data sets can provide you with an understanding of the ROI of site availability and allow you to visualize the impact that performance improvements have on user behaviors. The first take away from the fusion of these data sets is that we can see what the performance impact is on a website when the level of users increases. Here the page performance jumps from around two seconds to around nine seconds when usage levels increase from 200 page views to 600 page views!
Having this understanding of the impact that usage levels have on our site performance allows us to not only understand the level of users we need to be able to support, but also tells us if the performance of the site is satisfactory at these levels (i.e., consistent with performance at non-peak user levels). In this case, a usage increase of 300% resulted in a page load time increase of 350%! A more resilient system, capable of scaling, would certainly take a performance hit with a 300% increase in usage, but such a system would keep performance closer to the norm (around two seconds) as more and more users access the site. Without the combination of these data sets, we would have an analytics data set telling us that we have an increase in users on our site, but not telling us what their experience is — and we would also have a performance monitoring data set telling us that our performance has degraded, but only giving us basic insight into why .
Analytics data is very good at displaying the user’s story when he or she is using a web application. However, it only tells us that story when users are using the web application. Performance monitoring data is great at displaying the quality of usage of the web application and filling in the gaps left behind by analytics. Since performance monitoring is done with frequency, it means that we’re always collecting data and painting a picture of what the site performance would look like if a customer was accessing it.
Comments are closed.