Forum Moderators: DixonJones
We are able to track sales vs where from for all the usual suspects, such as all the Google variations, Yahoo, Overture, etc.
But in looking over 90 days of logs, I see not one single conversion from AOL. This cannot be right, as we have many orders with AOL email addresses.
But more importantly, isn't you concern for conversions about how well the ads convert, and not where they are appearing?
But if you really want to track AOL, at their webmaster site, they publish well documented lists of their proxy IP addresses, so if really wanted to do the analysis...
Larry
For example, latest 60 day period shows that people came from a total of 1784 URL's. 908 total sales for a total of $467,000+.
Not a single one of those URL's that shows a sale is AOL. 4702 hits from AOL search, 9454 from MSN search. MSN shows 24 sales. (google was 98,000, 151 sales just for reference).
Based on my analysis over the past 6 month period about the same problem and reading some posts in this site, I believe it is AOL.
However, here is the disclaimer: It may depend on the analytics software that you use and how that tags users/sessions and tracks them.
Like you said, the customer that uses AOL, has his/her IP possibly changed during the course of one session. That throws my in-house analytics program regularly. I track users by IP and get a low conversion rate from AOL and a huge number of sales from cutomers without referrers.
I did a little bit of path analysis of users referred by AOL and found that a majority of them (unlike other SE referrers) had incomplete paths or atleast my analytics system couldn't track them further. That lead to what seemed like a reasonable conclusion that its AOL thats throwing me off.
That said, I started experimenting with a session cookie based tracking. Don't have too many results to report yet. Will keep you posted.
I think a good approach is to open a log by hand for a day on which a lot of purchases happened, and examine each purchase session looking for AOL referrals. For those purchases, did the cookie field stay intact through the session? Is the session referrer clearly AOL? Did the stats program count that referrer correctly as AOL? If yes to all the above, then the remaining question is whether the analytics program is really using the cookie instead of IP.
I wonder if that would show up as if "the IP has left the building"?
It is possible also that AOL needs a different "model" of session cookie, that is something I had not considered. We do have orders from people with AOL emails, so we are pretty certain it is just not tracking.
The only problem you should have is if someone has cookies disabled, with I've found to be an infinitesimal number of users.
Congrats on the sales though :)
