I think we all have a webtrends story to tell...but basically my problem now
is that we started using analog and when we run webtrends on the same data,
we get different results--sometimes big differences like 700,000 vs. 900,000
page views for a site. I was wondering if anyone had done a comparison and
identified the differences in how the two products work, to help justify a
decision to abandon webtrends.
For example, in the case above, I tracked down the problem to discoving that
analog discards all 302 redirects while webtrends appears to count them if
the referer fails. This could happen, for example, if you have a lot of
.shtml pages supporting a legacy system where the .shtml redirects stale
links to new dynamic pages AND you are using a caching product like
SpiderCache to cache your dynamic pages. What happens in the log files is
that you see an entry for a page like:
product_12345.shtml with a 302 code and then, in some cases, you see a hit
to showproduct.asp?productID=12345 with a code 200. This is the normal
behavior. But if the showproduct.asp page is cached externally or running on
another server, there will never be the followup request in your server log.
At least this is what I could figure out so far.
My conclusion is that analog is discarding all 302s (reasonable behavior)
but webtrends must "care" if the referer actually succeeds. If there is no
subsequent 200 from the same IP address, the original 302 is counted. That
seems a little too "smart" for the likes of webtrends, but it is the only
way I can explain my discrepancies. When I look at the file types report
from both WT and analog, I see many more .shtml types recorded by
webtrends...not quite a 1:1 of course, but that is how I started looking
into the .shtml files as the possible culprit.
Anyway, I just wondered if anyone had a document summarizing the differences
which would help me as I have to explain why if we use analog, it means
fewer page views which means that 18 people (the site managers) are going to
be coming after me!
Dave Atkins
------------------------------------------------------------------------
This is the analog-help mailing list. To unsubscribe from this
mailing list, send mail to [EMAIL PROTECTED]
with "unsubscribe" in the main BODY OF THE MESSAGE.
List archived at http://www.mail-archive.com/[email protected]/
------------------------------------------------------------------------
