We have two websites, A and B. A is tracked with Google Analytics.
Users arrive at site A and log in to site B from there. We have teasers for content on site A on the log out page of site B. So:
A -> B -> A
My goal is to track how many visitors respond to the teasers on site B and visit A.
I understand that after 30 minutes of inactivity a pageview starts a new visit.
Am I right in assuming that if the user is done on site B in less than 30 minutes his original visit to A is simply continued? If so, would all traffic source info be "carried over"? For example, the user arrives via Google on A, logs in to B, logs off, goes to A. Would I then see Google as the traffic source when I look at the traffic sources of the landing pages on A?
What would be a good way to accurately track clicks on the teasers? Tagging the links on B would be my preferred solution. We cannot install GA on B.
Many thanks for your help!
I decided to dig into this subject myself as I found a lot of contradictory info. Here's what I've found out so far:
- The Firefox plugin Firebug (to look at request headers) and its extension Firecookie are useful tools for research on this topic.
- GA's utmz cookie contains info on the referrer in its utmsr value.
- During the same session, referral info is not overwritten by a second referral. So if the user takes the following path X -> A -> B -> A, utmsr is X upon the second arrival on A (again, same session).
- Sidenote: utmsr is overwritten if the second arrival is coming from search/organic.
- When __utm.gif is requested, one of the GET variables of the request is utmr. utmr contains the page referrer, i.e. B in the path example above.
- GA seems to use utmz for referral reporting and not utmr. This is based on preliminary research. However, it would also explain why you sometimes see referring sites which don't have links to the specific page you're looking at.
- I'll do some more research and post it here.