Stats analysis

"no referral"

[joeduck]

My Urchin stats indicate a large number of "no referrals" and I'm wondering if others have a similar high percentage of these when using Urchin. It is reasonable to assume they include some search traffic that can't be resolved to an engine?

1.(no referral)166,230 58.30%
2.google 31,495 11.05%
3.yahoo 26,216 9.19%
4.msn 10,589 3.71%
5.msn 3,175 1.11%

[larryn]

Joe,

No referrer can mean lots of things:

- Someone typed in your URL directly into the browser
- Someone clicked on a link in an email client
- Its a robot/spider (most don't offer referrers)
- Someone has security setup to strip out the referrer info

What does that mean to you? First I'd make sure that bots are segregrated in your analysis, then look at the numbers again to see if they make sense.

For one site I manage:

Total Visitors 3,790  
Total Page Views 7,231 
Normal Visits: 
Direct Entry 773 70.7% 
Portals & Search Engines 294 26.9% 
Links  18 1.6% 
Browser Cache/Local 7 0.6% 
Total Arrivals 1,092 100.0% 
Hit & Run Visit: 1,328 
Spiders: 
26 Spiders Visited: 1,370 

Hope that helps.

Larry

[joeduck]

Thanks Larryn -

I don't think bots are segregated by Urchin in this case and that would explain a lot. I'll follow up on that angle though this is through Valueweb hosting so I can't control defaults of the program.

[topr8]

in our view the majority of 'no referer' hits are from users who have blocked referer headers (assuming the robots are being filtered out properly - which i assume a sophisticated product like urchin is doing)

this opinion is based on the fact that we track our visitors by cookie and write their every move to a db. we have a large percentage of first time visitors who buy yet there is no referer info (some are maybe email recommended links - but mostly they will be search or other link traffic) [of course those who trash cookies on exiting the browser will also show this way too, but i don't think a very high percentage of users do that in our sector, which is not geeky at all]

[Jon_King]

>>in our view the majority of 'no referer' hits are from users who have blocked referer headers

Yes indeed. Nothing you can do about it. More and more are doing it. It sure makes 'standard' log files less accurate for tracking doesn't it.

You have to manipulate the log data to your usefulness. A tracking url, i.e. yourdomain/keywordproduct/keywordhtml.htm?tacknum1 is the most accurate tracking method. Track each link yourself via a unique url. What link is clicked and the corresponding page visit cannot be disguised. Track it yourself, the days of just interpreting an 'unloaded' log file are over. IMHO

[joeduck]

Interesting and thanks to all for helpful info.

I'm reading that there is little faith in Urchin's ability to parse the data in a relevant way and wondering if I should use direct analysis of the log file? Now that G has aquired Urchin I'm thinking a lot more sites will use it.

[cornwall]

This is a perennial one. Have a look at

[webmasterworld.com...]

Its old, but its as relevant today as it was then

[todd_c]

Hi Joe,

Agree with everything that has been said...however...

Question for you... do you do a fair amount of promotion of your website. i.e. Do you or does your company have a consistant pr program, perhaps buy print ads, have a good name your field, etc...

My guess is yes, based on your high "no referral"

Just curious.

[joeduck]

Hi Todd -

Actually _no_ promotion offline - I don't believe in anything but online ads. It's an old site so I'm on some bookmarks as a travel resource but the high "no referrers" does not seem to jive with common sense that would suggest most traffic coming from Search engines.

[Jon_King]

Thank you Cornwall. A truly good read.