What is the best way to determine which Google domains are sending traffic?

Site I am looking at gets traffic from google.com google.de google.co.uk etc .  Analytics lumps these together as "google". GWMT  doesn't share this and server stats through awstats/webalizer doesn't seem to break it down.  Am I right in thinking I'm going to be playing grep for the rest of the day?

That's a good question and one guess would be to look at referrers? Not sure that will give you anything though.

We're doing quite a bit of multilingual stuff and usually google.de will send the German traffic, google.dk the Danish and so on. So looking at the content you might be able to figure out a little more?

Expensive.... but PowerGREP (if you have Windows) is great for stuff like that.

But you could load it into Excel

Or spend 12 quid on this http://www.apacheviewer.com/ (which I've never tried).

Wouldn't Google's https connection kill the referrer data?  I read recently that https to https referer data stays intact however -- I haven't verified.

Edit:  I haven't checked this in a while but it *does* look like http_referer data is being passed on with Google's https.

Could try it out: https://www.google.com/#q=env.pl

I imagine Google Web Analytics is simply matching the referral domain against some fairly generic strings. E.g. If the referring domain contains 'google' then add it to the pile. This behaviour can't be switched off, but it can be overridden with more specific matches. Unfortunately it looks like these matches have to be added to your tracking code on each page. Google Web Analytics won't store this list.

The API method for adding a search engine is here:

https://developers.google.com/analytics/devguides/collection/gajs/methods/gaJSApiSearchEngines?csw=1#_gat.GA_Tracker_._addOrganic

and Brian Clifton has done the work already if you're prepared to part with a little money:

http://www.advanced-web-metrics.com/blog/custom-search-engine-hack/