Overview
On popular websites that have a large number of returning vistors, Urchin reporting may show long Average Length of Session statistics when the Urchin Traffic Monitor (UTM) is utilized. In addition, the Length of Session histogram in the Tracking section of the reports may also show a large number of sessions that are in the 30+ minute session length graph.
The underlying reason for this is that the UTM tracks visitor sessions with a session cookie, and any hits that share the same session cookie are considered to be in the same session. However, the session cookie is not destroyed until the visitor exits the web browser or explicitly destroys the cookie. Therefore, it is possible that UTM-tracked sessions can be many days in length, even if there are long periods of time between clicks from the visitor.
The next major release of Urchin will include additional configuration options that should mitigate this issue, by allowing for a default idle session timeout threshold to be set.
Detailed Explanation and Example
Consider the case where the Urchin Traffic Monitor is installed on a busy site that has frequent return visitors, such as a news portal or stock quote site. Typical visitor behavior to sites such as this might be to open up a browser to the site, spend 10-15 minutes surfing the links of interest, and then move on to other sites without exiting the web browser. The next day, this same visitor goes back to the UTM-tracked site. Since the browser has not been closed since the last visit, the session cookie is still the same as the previous day - and hence from Urchin's viewpoint it is still the same session (albeit with a long idle period). If this behavior is consistent across a large cross section of the site's visitors, then this can significantly affect both the Average Length of Session calculation, as well as increasing the 30+ minute graph in the Length of Session Report. Additionally, it will tend to skew the total visitor count for the period downwards.
See Also