WebTrends Statistics

[bobetko]

I am getting report from webtrends for period 2/1/2006 - 2/12/2006 (last 12 days) and I am getting following numbers:
Visitors 104,763
Visits 149,537

Entry Pages: 109,696
Every Visit has exactly one entry page, so number of entry pages should be equal to number of visits? Why I am getting 40,000 less visits?

Path Analysis for my domain is telling me: 60,768 visits. Why?

I am getting confused. Is this webtrends error, or I didn't setup something correctly....

Data is so different that is not usable at all. For instance, on the path analysis I would get 2-3% visits from pages that don't have link to the page I am analysing. How is that possible.

Thanks, for any help.....

[cgrantski]

Some visits don't enter at a page - they enter at a pdf or graphic or something else.

[bobetko]

thanks

[Gen Disorder]

The "EntryPages" total can be fooled by several things visitors can have on their client machines to defeat Webtrends reporting "sessionID" which is used to track visitors. On some visitors, every page is an "entrypage" because they client machine of the visitor blocks the tracking session from Webtrends.

Question 1: Why I am getting 40,000 less visits?
Remember Webtrends is only a "trend" and not verifiable data. Basically it will never add up perfectly. (If you brow-beat Webtrends Support, they will explain that to you.) Privacy tools, Anonymous browsing tools, very high browser security settings, session "kill" settings and some spyware tracking tools will all defeat webtrends tracking tools.

Qestion 2: Path Analysis for my domain...Why?
Path Analysis is only as good as the "sessionid" from Webtrends. If that session is dumped so is the "path" it was tracking.

Question 3: 2-3% visits from pages that don't have link to the page I am analysing. How is that possible?
The visitor may be using a search engine to find things on your site. They bounce in and out of your site using multiple browser windows that give the appearence of hyperlinking to non-connected pages. Webtrends uses the HTTP-Header object called "REFERER" to assemble path and that can be fooled, especially on Mozilla-based browsers that pass HTTP objects back and forth between browser windows - This passing of objects between browser windows makes "logins" and other user sessions work in new browser windows. Otherwise the user would have to login to the same website repeatedly for each new window.

Cpt. Kaos
Who loves you, baby!