Summary.Net Archives
 
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Summary-Talk] 3.0b5 feedback - visits vs pages



Hi Jason,

I have been testing 3.0b5 side-by-side with 2.6.8 with identical
configs and so far it's been very good.  Lots of great new features.

However, I do have a couple questions. Specifically, I'm looking at
"Referring Domains" report (/report/20) and within that the columns
for Visits and Pages.

Having restricted my log processing to a single day, I am seeing the
following results for a certain referrer:

version		2.6.8		3.0b5
Visits		283		148
Pages		267		145

I have manually counted the referrers for this certain domain in the
single logfile being processed, and I come up with:

Total matching requests with this referrer and 200 response: 285
Total matching requests with this referrer, 200 response and unique
IP: 188
Total matching requests with this referrer and 301 response: 40
Total matching requests with this referrer resolving to proxy.aol.com: 7

I'm assuming 301's are not counted as "pages" but do contribute to
the visit count if they are the only hit in the visit.

So my questions are:

1) Are visits being reported correctly in either version?  My testing
implies that visits may be over-reported by ~50% in 2.x and under-
reported by ~23% in 3.x, unless I'm missing some of your criteria for
defining a visit beyond what's in your Glossary.

2) Are pages being reported correctly in 3.0?  The reported value
seems to be off by ~50% based on my manual calculations.

3) Should the "visits" column always be equal to the sum of the
visits displayed when clicking on the domain drill-down?  In this
case, the sum of the selected referrer visits is 286, compared to 148
in the parent report.


I love the new (Tag xxxx) entries in this report, as they are quite
helpful.  They would be even more useful if their drill-down links
worked.  I'm not sure if this is a technical issue or just an
oversight in implementation, hopefully the latter.

Please let me know if you need any more details or samples to
reproduce this problem.

Keep up the great work!

Thanks,

Dale
-------------
Go to <http://summary.net/list.html> to update subscription info.