homepage Welcome to WebmasterWorld Guest from 54.161.175.231
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member

Home / Forums Index / WebmasterWorld / Website Analytics - Tracking and Logging
Forum Library, Charter, Moderators: Receptional & mademetop

Website Analytics - Tracking and Logging Forum

    
RSS and Safari and polluting Analytics stats
Inflating referrals and lowering avg pageviews
caribguy




msg:4191160
 10:26 pm on Aug 23, 2010 (gmt 0)

I recently ran into a bunch of 'phantom' hits for one of the least trafficked pages on my site, apparently resulting from a very targeted Google query.

Since the average no. of page views for this series of visits was close to 1.0 and resulted from a single ISP, I decided to investigate further: fearing an attempt to Googlebomb that page.

Upon checking that specific query in my Apache logs, I only found the initial visit that was referred from the Google query. Here is a sample of visits:

"GET /folder/file HTTP/1.1" 200 16780 "http://www.google.com/search?[query removed]" "Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_8; en-us) AppleWebKit/531.21.8 (KHTML, like Gecko) Version/4.0.4 Safari/531.21.10"
"GET /folder/file HTTP/1.1" 200 16777 "http://www.example.com/folder/page2" "Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_8; en-us) AppleWebKit/531.21.8 (KHTML, like Gecko) Version/4.0.4 Safari/531.21.10"
"GET /folder/file HTTP/1.1" 200 68451 "-" "Mozilla/5.0 (Macintosh; Intel Mac OS X) Word/12.23.0"
"GET /folder/file HTTP/1.1" 200 16780 "-" "Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_8; en-us) AppleWebKit/531.21.8 (KHTML, like Gecko) Version/4.0.4 Safari/531.21.10"
"GET /RSS/feed HTTP/1.1" 200 2429 "-" "Apple-PubSub/65.11"
"GET /RSS/feed HTTP/1.1" 200 2426 "-" "Apple-PubSub/65.11"
"GET /RSS/feed HTTP/1.1" 200 2429 "-" "Apple-PubSub/65.11"
"GET /folder/file HTTP/1.1" 200 16780 "-" "Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_8; en-us) AppleWebKit/531.21.8 (KHTML, like Gecko) Version/4.0.4 Safari/531.21.10"
"GET /folder/file HTTP/1.1" 200 16795 "-" "Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5_8; en-us) AppleWebKit/531.21.8 (KHTML, like Gecko) Version/4.0.4 Safari/531.21.10"


What this looks like to me, is that GA is tracking each subsequent visit from this particular user as resulting from their initial Google search. This is not in line with reality and inconsistent when comparing to other RSS feed users.

 

caribguy




msg:4192622
 4:28 pm on Aug 26, 2010 (gmt 0)

Maybe it's just my imagination... Anyone else?

caribguy




msg:4201443
 12:21 am on Sep 14, 2010 (gmt 0)

Looks like I am not alone after all. While investigating a separate issue today, I found this unusual header in my logs:

HTTP_X_PURPOSE 'preview'

The issue identified in my first post may be related to Safari's Top Sites feature, which displays live thumbnails of frequently visited pages. Not sure, help and insights much appreciated.

An article about this lovely feature here: [sunpig.com...] (mods: if there is a better reference please replace)

Edit: there may be separate issues at play here, both related to Safari and polluting Analytics stats.

Receptional




msg:4201532
 8:04 am on Sep 14, 2010 (gmt 0)

I'm not silent because I'm ignoring you. This is interesting and needs answering. I wish I had the time to figure it out! So I hope someone else will.

caribguy




msg:4201789
 8:06 pm on Sep 14, 2010 (gmt 0)

I hope so too, and obviously I wish I had never discovered this situation since it takes time away from more productive work :)

So far, I have uncovered two occasions -about 100 and 150 visits each within a one week period- where this effect showed up (I think the miscounts have distinct causes). There may be more occurrences, as the effect is harder to see for pages with higher visitor counts.

When I have more time on my hands, I'll get a copy of Safari and try to replicate the problem. For now, I think that it happens when a Safari user

a) visits a particular page (or site?) a few times within a short period, and it gets added to Top Sites - I suspect this could also be done by pinning it there manually, like in Chrome

b) subscribes to a rss feed on the site

In both situations, Safari most probably resubmits the original credentials (including cookies, session variables, etc) that it originally used when the visitor first accessed the resource. Each time Top Sites refreshes the content or when the rss feed is viewed, the call to GA is made with the original referral and other tracking information still in place.

Google could expire the cookie (this will never happen, hah!) or use the original access time to filter out these duplicate calls (or count them differently).

I have noticed that for the past 50,000 visits or so, Safari accounts for 13% of visits, but only 10% of page views. This seems to be a significant difference.

In my scenario, I expect single pages to be added to bookmarks, top sites and feeds fairly often. We actually provide the very same feature (a list of favorite pages) to our registered users.

One option to look at is to not serve the analytics code to those visitors, or to add a pageTracker._trackPageview(tracker_id); tag.

Maybe some of you with the right connections could have the powers to be (Avinash and friends) investigate the issue?

roodle




msg:4237186
 2:47 am on Dec 1, 2010 (gmt 0)

caribguy, did you get to the bottom of this? I've recently been seeing something similar but on a smaller scale. Was it Safari preview?

Global Options:
 top home search open messages active posts  
 

Home / Forums Index / WebmasterWorld / Website Analytics - Tracking and Logging
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved