If you get Google Images traffic then Google Analytics is not giving you the correct number of visits… For example, looking at a mainstream site I work on… Google Analytics said it had 2.977 million visits last year and 9.41 million page views. Google Analytics says 39% of the visits and 31% of the pageviews came from Google Images…
WRONG… Our client’s corporate standard for analytics is Unica, and Unica says there were 4.587 million visits and 11.91 million page views.
The difference is that Google Analytics doesn’t count hits on framed pages but Unica does count framed pages. Google Images frames your page - so the two stat packages report (vastly) different numbers. It’s not until the person clicks something to remove the Google Images frameset that Google Analytics counts the page hit / visit - even though the person is seeing your entire page before then…
Bottom line - the real amount of traffic that you get from Google Images is more than double what Google Analytics says it is… If you get a large amount of Google Images traffic then you’re numbers can be way off from what you see in Google Analytics.
I only quite recently started to use Google Images myself, it’s actually very good but I don’t like the way it works when you click on the images.
I can also see how that could be counted in different ways through analytics or server stats. Like you mentioned I’m sure I would find the same if I compare my server stats with my analytics account. After all even if it’s a framed view it’s still a visit / hit on the website. I wonder if they count it as “their” visitor while it’s in framed view.
On smaller websites it probably doesn’t matter but once you have a larger volume of traffic it will make a major difference on load and usage. They really should include it in their stats otherwise it’s simply not accurate.