Forum Moderators: DixonJones
Thanks for any help!
Oh, in case it helps - - using WebTrends Reporting Center - eBusiness Edition
Version: 6.1b
Platform: win32
Build Number: 7601
I don't use WebTrends myself - I'm sure there are others here who do and may have something to add to this - but it sounds like visits that are reported as zero-length are from people who come to your website then leave immediately.
If it's the "Visits by Number of Pages Viewed" report, and you're referring to the line indicating 0 pages viewed, then you're looking at activity for which no actual PAGES happened. For example, if somebody's site shows one of your images using a link to your server, then any view of that page on the other site will hit that image file on your server, resulting in a "visit" of zero pages. Or if somebody's site has a link to a downloadable pdf or an executable on your site, same thing, as long as pdfs and exes are defined (in WebTrends) as not being "pages."
Is this by any chance the answer to your question?
WebTrends determines the session lenght by looking at the time of the first request and the last request. If there's only one request it _could_ result in a zero lenght visit.
Zero-page view visits refers to hits that were not a valid HTML or script page. This would include GIFs, JPGs, etc. so I believe WRC is referring to those in your reports.
When determining the actual visits for my clients, I always take the total visits and minus this number for actual "visits". Often, the numbers can be quite interesting, esp. if you have banners, images showing up in Google Image search results and so on. So the benefits in some reports (and I'm not looking at WRC as I write this!) is likely to show trends from these types of visits - e.g. a marketing campaign you would obviously want to show those banner visits to determine impressions.
As for the documentation, I agree, WebTrend's docs are relatively good at explaining the product screen-by-screen but beyond that, it's trial-and-error as a user!
Reports I'm looking at are "Visit Summary" report table where the data does not include zero-length visits. I'm also looking at the "Visitor Trends" report table, where the data DOES include zero-length visits. In both table reports I'm looking at 'Average Visit Length' - and in my reports they are always different. For example, these are from a report :
Visit Summary \ Average Visit Length = 00:14:08
Visitors Trend \ Avg Visit Length = 00:4:15
Bit of a difference.
And the info on this in NetIQ Article NETIQQKB27110 enlightens with "A zero length visit occurs when all hits in that visit are logged with the exact same time stamp". That I don't understand. And again, why include it in some tables (Visit Summary) and not in others (Visitors Trend)?
Perhaps my mistake is in even thinking that these are comparable data?!?
Thanks again!
"A zero length visit occurs when all hits in that visit are logged with the exact same time stamp". That I don't understand
Log example:
192.168.100.100 - [01/Feb/2004:11:52:30 +0100] GET /index.htm
192.168.100.100 - [01/Feb/2004:11:52:30 +0100] GET /picture1.jpg
192.168.100.100 - [01/Feb/2004:11:52:30 +0100] GET /picture2.jpg
In the above example all three files are requested at excactly the same time. Therefore WebTrends doesn't have a "session start" and "session end" time. Hence it's a zero length visit.
So when you are trying to build a picture of how long a user is actually staying on the site, what data are you using?
how long a user is actually staying on the site
To be honest I'm not so interested in this number. Are long session times good or bad? A long session could be a result of the user not being able to find the info he needs - or a result of the user finding the info on the site valuable. You don't know... It's impossible to conclude (almost) anything based on visit length.
And if you can't make any conclusions then why bother looking at the numbers ;-)
Instead I would focus my energy on eg. single page visits a.o. - this gives me a much better picture on my ability to "catch" the users.
A visit of two pages views has a duration equal to the difference between the time of the first page request and the time of the second page request – in other words, it’s just the time spent looking at the first page. The second page viewing time is unknown. So somebody viewing page A for 30 seconds and page B for 30 seconds, then leaving the site, is recorded as a visit of 30 seconds.
But the line in the Visits Summary report labeled “0:00” actually includes all visits 59 seconds in length or shorter – the rows are labeled with the starting point of the time period. So these aren’t really visits of length 0:00, although some of them could be.
The numbers in Visits Summary and Visitors Trend are different because the denominators are different. Both divide the total viewing time by the number of visits, but number of visits is defined differently. The Visitors Trend report includes all visits including single page visits. The Visits Summary report includes only visits of more than one page.
True, the labeling for these in WebTrends is pretty crummy. But once you know this, you can interpret them like this:
Visits Summary “average visit length” is an approximation of the average visit length for visits whose length is actually known – i.e. visits that had two or more page views.
Visitors Trend “average visit length” is an approximation of the average for all visits.
As I understand it, this inconsistency and/or lack of clarity will be taken care of in the new release happening in the next month or two.
