Forum Moderators: DixonJones

Message Too Old, No Replies

Tracking named anchor (#) in URL

Any way to get Webtrends Reporting Center to track it?

         

zgb999

4:31 pm on Jun 29, 2004 (gmt 0)

10+ Year Member



A named anchor is by default not a page of itself but an anchor on a page (e.g. page.htm#anchor).

I could not find a way to get Webtrends Reporting Center to the point where it shows me how often a named anchor was used. page.htm#anchor is stored in the logfiles but how can I get a report on it?

cgrantski

7:47 pm on Jun 29, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



In WebTrends you could get around this if you use WT's URL Search and Replace function. Tell it to change the "#" to "?anchor=". With this approach, you could work with parameter analysis reports, or tell WT to display the parameters as part of the URL which will make the whole thing display in page reports.

Or, change ".htm#" to "-" resulting in a URL of "page-anchorname" rather than "page.htm#anchorname. WebTrends should be able to handle files with no extension (depends on the version of WT).

There are several ways to do it, as you can see, and your choice will depend on what's happening with the rest of your logs. But they all start with URL Search And Replace. (On some sites I deal with, every single page on the site gets URL S&R'd - it's a great feature.)

One important thing to know about URL S&R is that it happens before anything else happens at all in WebTrends. It's the very first operation. (knowing that will help you make other decisions that may come up).

And, URL S&R does not change the original log at all.

Hope this helps.

zgb999

9:13 am on Jun 30, 2004 (gmt 0)

10+ Year Member



Thank you very much!

The experience within WW is just great!

Unfortunately I could not find how to replace a single character.

I have

Replace from:
start of first: #

Up to:
start of next:

But I could not find a wildcard to tell Webtrends to change only the #.

The dropdown for "Up to" gives me only 3 options:
start of next
end of next
end of string

Only the "start of next" option would work but there might be any caracter after the #. So I would need to tell Webtrends that it shall change up to start of next * [wildcard, any caracter].

How could you solve this?

I really appreciate your help on this!

cgrantski

4:22 pm on Jun 30, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I'm pretty sure that what I've done there is have the second choice be "end of next" with "#". Try it on one day's log or something.

dcrombie

9:34 am on Jul 1, 2004 (gmt 0)



AFAIK anchors don't appear in the server logs b/c they are handled internally by the browser. That would make it hard to generate a report ;)

edit_g

9:44 am on Jul 1, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Yup - I think once the server has washed its hands of the page by loading it fully, it becomes a browser thing unless the page is reloaded or navigated away from.

You could make the link something like [example.com#whatever?anchor=faq1...] - this would reload the page and let you know which link they're clicking. It would also reload the page, which defeats the point a little... ;)

cgrantski

12:02 pm on Jul 1, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



The originator of the topic says they are in the logs. It's possible to get it to happen without the visitor experiencing anything. It's a lot like logging individual segments within a flash movie.

dcrombie

12:35 pm on Jul 1, 2004 (gmt 0)



The anchor is only ever passed in the 'referer' string - never in the request.
Read the posts from killroy in this thread:

[webmasterworld.com...]

From our logs it looks like the only browser consistently passing the anchor in the referer is MSIE5.x(Mac_PowerPC). That suggests to me that it's not something that should be done ;)

cgrantski

1:37 pm on Jul 1, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Excellent point. I assumed ZGB was seeing it in the URL-requested field because they'd used a trick to get it there (flash movie segment logging etc stuff). ZGB - what field is the # in?

zgb999

7:00 pm on Jul 1, 2004 (gmt 0)

10+ Year Member



Sorry for the late answer. Here some random examples of the logs:

217.232.254.152 - - [16/Jun/2004:17:23:24 +0200] "GET /pageXY.htm HTTP/1.1" 304 - "http://www.domain.com/pageXY.htm#02_09_Internet_Widget" "Mozilla/4.0 (compatible; MSIE 5.23; Mac_PowerPC)"

80.140.162.115 - - [07/Jun/2004:11:17:04 +0200] "GET /pictures/news/0302-picture.gif HTTP/1.1" 200 12994 "http://www.domain.com/pageZZZ.htm#03_02_text" "Mozilla/5.0 (Windows; U; Windows NT 5.1; de-DE; rv:1.0.1) Gecko/20020823 Netscape/7.0"

dcrombie

9:54 am on Jul 2, 2004 (gmt 0)



That matches what I see - the following browsers pass the anchor as part of the referer string:

MSIE 5.x (Mac)
Netscape 4.7x
Netscape/7.0x (Mozilla rv:0.9.x - 1.0.x)
Safari/85.7 (Mac)

In total they add up to <<1% of browsers so I don't see any value in analysing the anchors for such a small sample.

cgrantski

11:21 am on Jul 2, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



dcrombie, is there a place to go for that kind of detailed information on browser behavior? A source for such info would be great to know about. :-)

Going back to the original question, now that I realize I was working from an entirely incorrect premise ... it would take a not-too-difficult text manipulating script to get the bookmark into the URL as a parameter, or perhaps the referrers report output could just be analyzed ... but I agree that this group of browsers is really small and probably doesn't accurately represent the whole visitor population anyway.

dcrombie

11:36 am on Jul 2, 2004 (gmt 0)



I'm only going from what I see in our logs - it seems that the browser-writers included the referer-anchor in their early versions, then 'wised up' and stopped passing it. There must be a standard somewhere but I haven't been able to find it.

It would be GREAT if there was this kind of information on the WWW but I guess the programmers work on the premise that any problems are going to be fixed 'in the next version'.