This document provides a brief overview of the differences in the ways that Urchin 3 and Urchin 4 report visitor sessions. This may be helpful to Urchin 3 customers who are evaluating Urchin 4 or have upgraded to Urchin 4 and wish to correlate report statistics between Urchin 3 and Urchin 4. For a more detailed description of visitor identification in Urchin 4, please see the whitepaper entitled Visitor and Session Identification Overview available on help.urchin.com.
In Urchin 3, the term visitors refers to the count of distinct visitor sessions to a website, as determined by the IP address of visitor. Urchin 3 uses strictly IP-based tracking to determine visitors, it does not examine cookies, session IDs, or browser/platform information to further differentiate between visitors. Nor does it attempt to report on unique visitors - multiple visits from the same IP address will each cause the visitor count to be incremented for each visit.
Urchin 4 has several different tracking methodologies, with the default being IP+UserAgent (browser/platform combination). Like Urchin 3, this methodology determines visitor sessions based on the IP address of the visitor. However, Urchin 4 also has some key differences that provide for more accurate reporting of sessions in the default configuration.
- A visitor sessions must have at least one valid hit
- A visitor session must have at least one pageview
- Session data is maintained across Urchin processing runs
- Browser and operating system information for the visitor are examined (i.e. the UserAgent string) to further differentiate between discrete visitors who have the same IP address.
- Special algorithms to handle AOL visitors
If it is desirable to configure an Urchin 4 profile to provide reporting that most closely matches the numbers produced by Urchin 3, the Visitor Tracking Options in the Advanced Settings for the profile should be configured with:
- Visitor Tracking Method set to IP Only
- Session Requires Pageview set to off
Remarks
- For visitors that reside behind a firewall or that use a web proxy, the hits logged by the viewed site all appear to come from a single IP address. This causes IP-only session tracking (e.g. Urchin 3) to under-report the number of actual sessions from these visitors, as the discrete sessions appear to be all one long session from that single IP address. Urchin 4 is more accurate because it can differentiate between visitors behind a firewall/proxy if those visitors have different browser and/or operating system versions. This is the default behavior for Urchin 4, although selecting IP-Only for the Visitor Tracking Method of a profile reverts Urchin 4 back to Urchin 3's IP-only tracking logic.
- Accurate IP-based tracking of AOL visitors is rather difficult due to the proxying behavior that AOL has implemented. The net effect is that a single session for an AOL visitor will typically be logged with many different IP addresses - worst case, a different IP address with every click. For IP-based tracking, the resulting reports will show a separate session from each click, greatly inflating the AOL and overall session counts. Urchin 4 has special algorithms to detect and compensate for the AOL proxying behavior, and can accurately consolidate hits into discrete visitor sessions. This has the net effect of considerably reducing the number of AOL sessions compared to Urchin 3's reporting. This is the default behavior for Urchin 4, although selecting IP-Only for the Visitor Tracking Method of a profile reverts Urchin 4 back to Urchin 3's IP-only tracking logic.
- Overall session counts as reported by Urchin 4 may be less than Urchin 3 since a session in Urchin 4 requires a valid hit, e.g. a pageview, image or other content was delivered by the webserver. In Urchin 3, a session was recorded even if it contained only hit(s) with 400 or 500 status codes, such as 404 (Page not Found).
- Pageview counts in Urchin 4 will typically be less than Urchin 3 since Urchin 4 does not treat .css and .js files as pageviews, whereas in Urchin 3 these hits were treated as pageviews.