Forum Moderators: martinibuster
I've run an analysis, excluding search engine spiders and our own accesses, based on a dataset of roughly 32,000 unique visitors (60k page reads and 330k hits).
The geographic spread of users is:
US: 53%
UK: 18%
Canada: 7%
Australia: 4%
India: 2%
The breakdown of visitors arriving from each of the search engines is rather Google-heavy for our main site:
Google: 85%
Yahoo: 10%
MSN: 4%
AOL: 1%
There were visitors from 87 different Google domains, the top 5 of which were:
google.com: 60%
google.co.uk: 18%
google.ca: 7%
google.com.au: 4%
google.co.in: 3%
Now to the punch line:
Of the visitors from google.co.uk, 26% selected the "search from the UK option".
Expressed in terms of all Google visitors, just under 5% selected the option to search UK pages only.
In terms of Adsense income, our site is primarily a business site, and we track income during the day, so we can guesstimate where our advertising revenue comes from. Income from the main sources is estimated at:
US/Canada: 45%
UK: 37%
Australia/India: 18%
As far as I can tell, "no". For three reasons.
Firstly, there are 6 aol domain referrers to our site, 3 of which appear to be aol.com, and one of which is .co.uk. The breakdown of visitors from these is:
* aol.com: 67%
* aol.co.uk: 30%
* aol.ca: 1%
* aol.fr: 1%
* (rounding error accounts for the missing 1%)
This pattern seems to echo the pattern of visitors reflected elsewhere in site stats.
Secondly, if you look at the original post you'll see that the total referrals from aol is very small. So, even if many of the aol.com users were actually based in the UK going via a US server, they wouldn't be able to distort the stats much.
Thirdly, our hourly site accesses increase significantly when US business hours come into effect, reinforcing the idea that the majority of our visitors are from the US.
Wow, thanks...now I have to try and digest these over lunch...well, a 5 minute sandwich:-(
I was asking about the geographical spread rather than referrers - ie AOL members rather than those who specifically find you via AOL search.
The reason I ask is that I was once on AOL and in the UK. With a brand new site, I was the only visitor (building it, checking it, etc) and the country breakdown of visitors read 100% USA although I was in the UK. So I assumed from that AOL users the world over are deemed to be in the USA as far as the stats were concerned.
Joe
and the country breakdown of visitors read 100% USA although I was in the UK.
I've noticed that too. I constructed a site recently and I was the only visitor and it said I was from the USA.
Why?
My broadband is through pipex.com...it even gave the pipex reference and they are as Brit as they come I would've thought, therefore that analysis, Awstats, is plainly incorrect with the country of origin.
That's a good question.
The way in which our stats software calculates the geographical spread is by working back from the visitors IP address in the log file - looking up the IP address in a database that associates countries with IP addresses.
This database is not guaranteed to be accurate, but the three reasons I cited - two about referrers and one about hourly access - seem to corroborate the host-based geographical analysis, hence I think it is probably fairly accurate.
If your stats analysis did not distinguish between UK and US based hosts, it might be that the software wasn't using such a sophisticated method of working out where your visitors came from. Alternatively, it might have been that you were going through a US host, which would be difficult to pick up even in our analysis. However, because we are a business site, and the hourly access statistics seem to suggest that visitors are mainly from the US, I doubt that this is happening on any large scale.
I can put our software to the test in a similar way. I'll try an unused site and let you know later what happens.
My stats do distinguish between UK and US visitors, but from my own experience AOL customers are classified as US visitors.
I've run an analysis using a wildcard filter to include all aol's proxy servers, and no others. All the AOL visitors in this analysis are reported as from the geographic region "unknown". I could believe, however, that some analysis packages might report them as US as, from a visual check, all the servers have aol.com in the domain name.
The geographic stats I reported earlier included all hosts, whether the country of origin was known or unknown. "Unknown" reported in its own category, and scored at 2% (a similar level to India).
What this suggests to me is that I get very few visitors from AOL (as reflected in the referrer data) and that the geographic data isn't significantly distorted by AOL.
Hmmm... in our logs, over 90% of visitors coming through a domain with "pipex" in it are reported as being from Spain. The rest are classified as "unknown".
Do pipex house their servers in Spain nowadays?
