|WebTrends Session Tracking - Cookie vs IP address|
| 11:54 am on Jul 13, 2009 (gmt 0)|
I have two questions regarding WebTrends Session Tracking mechanism.
First one is, how exactly First Party Cookie Session Tracking option works?
How can WT track visitors by cookie, if no cookie is issued? I'm puzzled since there is no cookie info inside my IIS log files. When First Party Cookie tracking option is activated, WT shows 91 visits from 56 visitors. When I switch session tracking to IP/User Agent option, WT shows 163 visits from 101 visitor. What is the catch here?
The second question is about SDC mechanism. I understand that website's performance will not significantly decrease with additional server call to SDC server, but what happens when SDC server is down? Will the web page be idle while it is waiting for a response from SDC, for some amount of time?
Thanks in advance for answers.
| 5:59 pm on Jul 21, 2009 (gmt 0)|
In the rare event that the SDC is not available, it will not affect the page load time if the code is placed at the end of your page. (You may have an issue if you place the code incorrectly in your HEAD tag)
| 7:18 am on Jul 22, 2009 (gmt 0)|
Thanks for the answer, unilytics.
About Session tracking, I think I figured it out. If you select First Party Cookie option for session tracking, even if you don't have SDC and you don't have any cookie information in your web server's log files, WebTrends will automatically switch to session tracking by IP address. The only thing that remains unclear to me is why I have different results when I select First Party cookie or IP/User Agent options, when both methods should use IP address for identification.
Now, about SDC server, yes I will place JS tag at the near end of the page, just above the closing </body> tag, but that is still not the end of the page. So, theoretically, page will be loaded like 99% of it, not 100%. It will stuck at JS code where a call to SDC is made and it will wait for SDC responce. Again, this is just a theory. I thought someone with practical experience can confirm this.
| 10:24 pm on Jul 22, 2009 (gmt 0)|
The reason you will see differences between "FP cookie" and IP/UA options is:
-different users on the same public IP can share the same IP/UA
-only cookies can make the distinction between those users.
To answer your other question, if your website is constructed with tables and the code is still inside the table, it could delay the overall page load. If you do place the code at the end of the page, it would have the equivalent effect as a broken image at the bottom of your page. (browsers vary on how they display "load" information for the page, but most modern browsers will render the entire page in full while it waits for the broken image)
| 7:38 am on Jul 23, 2009 (gmt 0)|
Thanks again for the answer, I appreciate it.
Things are clearer to me now.
| 12:02 am on Jul 25, 2009 (gmt 0)|
Deja, I'm still puzzled. From other posts of yours, I know that you sometimes analyze log files and sometimes SDC logs. From what you said at the beginning of this topic, I'm guessing that in the situation you are describing, you are running WebTrends on some server logs. You're not talking about SDC logs (which WILL have a cookie in them). You said the server logs have no cookie in them.
At the same time, you're saying "When First Party Cookie tracking option is activated, WT shows 91 visits from 56 visitors. When I switch session tracking to IP/User Agent option, WT shows 163 visits from 101 visitors."
If it's true that you are analyzing the same cookie-free IIS logs both times, and all you're changing is the Session Tracking option, then WebTrends should be doing the exact same thing both times, i.e. using IP/UA sessionizing. Your results should be exactly the same. But they're not.
Am I reading what you're saying correctly? If so, something else is going on. Maybe you've uncovered something odd about WebTrends that I didn't previously know about, so I'd like to pursue it.
Are the page view numbers the same both times?
| 1:58 pm on Jul 27, 2009 (gmt 0)|
You understood correctly my post.
The thing is that I have operational SDC and it's logs, but in test environment.
Now we want to utilize SDC in our production environment, and before I started
installation of SDC I took a look at existing WT profile which uses IIS log files,
and saw that it uses First-Party Cookie as session tracking method, even with no cookie information is log files.
I tried just now to reproduce again the issue - I re-analyzed the same log file with different session tracking methods,
once with First-Party Cookie and once with IP/User Agent, but now I got the same results - same number of visitors, visits and page views.
I can't reproduce those different numbers from my previous post, so let's say that I didn't perform the steps correctly.
| 12:54 pm on Jul 28, 2009 (gmt 0)|
I do that all the time - was I looking at the wrong time range? Did I change a setting? It's both humbling and reassuring that the problem is almost always with the human, not the computer. And with a tool like WebTrends that has so many choices, it's doubly likely to be the human.