Back-to-School Retail Web Site Performance Analysis

It’s somehow time to mark the end of summer with Labor Day weekend. And with that time, comes the rush we remember well as Back to School season. As most students are already back or soon to return to school, we decided to look at some major retailers and how their back-to-school special websites performed during the season.

As you can see from the graphic above, the performance of some of these sites spanned a wide range from six seconds average End User Response Time to over 17 seconds.

The wide range can at least in part be attributed to different approaches to design philosophies taken by the retailers. In some cases, a retailer may choose to design a page with too many graphics for individual products or specials. Those cause the size of the page (in terms of data that needs to be downloaded) to be comparatively large, whereas others may take a more streamlined approach to fewer graphics on the main page.

The problem becomes that companies make these types of decisions without any idea of how it affects their site performance. The impact of that site performance then reflects on the conversion rate of their web page.

In addition to the average response time, another interesting aspect of performance is the variation over time. Again, the graphic at the top of the article shows a wide disparity where some retailers have relatively consistent performance within a fairly narrow range, whereas other retailers have a much wider range of performance over time.

As a general rule, companies should certainly strive to provide customers with as consistent an experience as possible, and performance is a key part of providing a consistent experience.

The advantage of measuring site performance with AppDynamics Browser Synthetic Monitoring (Beta) is that they are consistently taken from our cloud-based geo-distributed intelligent agents. This provides a consistent and repeatable baseline measurement without all of the vagaries that can be introduced from real-user requests.

A company should strive to understand why their site may be providing such widely different response times and seek to reduce the range of performance to a more consistent and narrow range.