Jeremy Wadsack wrote:

Tim Larson <[EMAIL PROTECTED]> (Wednesday, October 01, 2003 12:02 PM):
1. I (meaning the marketing folks) would like to get a breakdown of
search engines by query (where are people looking for ProductX coming from?), and also a breakdown of queries by search engine (most Googlers are looking for which product?).

You can do this by specifying individual search terms or engines with a REFINCLUDE command and generating a single report for each item you want to inspect. See http://analog.cx/docs/include.html for details on the command.

See http://analog.cx/docs/faq.html#faq128 for reasons why this isn't
available as a single report.

I guess in order to keep analog's report generation fast, I can do a little extra work myself. :) However, it does mean that I may be running analog a couple dozen times, so I might end up spending just as much time crunching data as I would have if analog did all the cross-indexing. :( And now the results are spread amongst multiple report pages, which isn't as convenient.


If this is "the final word" on the matter, it would potentially be useful if analog wrote an intermediate results file (like AWStats does) that all subsequent reports would get generated from. My site is relatively small, and I suppose I could build all the reports I want in under ten minutes. But surely there are others with much larger sites that would benefit from separate data analysis and report creation steps.

2. I (meaning my "stat geek" alter ego) would like to get a breakdown of
browser by OS (which browsers are Windows people using?), and also OS by browser (are Netscape people mostly Windows or Mac users?).

Same as above. Except use the BROWINCLUDE command with appropriate patterns. You might need to check the source code for patterns used for these reports (especially OS).

One shouldn't need to have the knowledge to read source code just to get a report that seems so obvious to webmasters. Storing these patterns in a text file would make it a little simpler, and might allow for more flexible customization. As an added benefit, I could set up a nightly job to get the latest definitive browser list, so my stats accurately report the newest browsers.


3.  How can I keep robots off my browser report/summary?  I'd like to be
able to optionally lump them together (Known robots) like they are in the
OS report, or exclude them altogether.

You can create BROWEXCLUDE commands that match the BROWINCLUDE patterns in your config file. There's also a more comprehensive list of robots linked on http://analog.cx/helpers/#conffiles.

I'm already using one of the comprehensive robot lists. Since they're _already_ defined as robots in the config, a simple "BROWEXCLUDE robots" seems obvious. Why should I have to define all of them twice, and perhaps miss updating one of the lists properly?


Why the decision to lump robots together in the OS report but not the browser summary? Why not a "OSCOMBINE robots" option, and a "BROWCOMBINE robots" option as well? I know the platform Google (or anybody else) runs their spider from is irrelevant, so that's a good reason to drop it from the OS report, but from the POV of a webmaster using the reports I want a summary of what _humans_ are using to browse with. Maybe it's interesting to know that .82% of requests were from spiders, but if I want more detail than that I'll look at the full browser report instead.

I hope this doesn't appear argumentative. I'm just wondering what the design assumption here is. It seems to shove extra work on the analog user to adapt, when analog could be doing something more sensible by default. Or at least as a config option. Options are good, and make analog a better tool. Maybe some of these ideas are already in the works, but I haven't seen a "planned features" page.

Tim

--
Tim Larson, Webmaster, LovelySkin.com
For a more beautiful you, surf us!
402.334.SKIN

+------------------------------------------------------------------------
|  TO UNSUBSCRIBE from this list:
|    http://lists.isite.net/listgate/analog-help/unsubscribe.html
|
|  Digest version: http://lists.isite.net/listgate/analog-help-digest/
|  Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
|  List archives:  http://www.analog.cx/docs/mailing.html#listarchives
+------------------------------------------------------------------------

Reply via email to