Robert_Charlton - 8:11 am on Mar 3, 2012 (gmt 0)
So I checked from our London based VPS to make sure and the keywords are not on the first page, they are where they were before. We are based North-West UK, approx 300+ miles away from London.
Location has been a key factor in Google search results for a while, since early after Place pages were introduced to the first page.
I've noticed that even when you're signed out with cookies cleared, or when you're signed in with search history turned off (and note that you must be signed in to disable search history, because search history settings are "remembered" via cookies)... geo-location is the big variable which still comes into play, no matter what your personalization settings are. Pretty obviously IP and ISP based.
You apparently can manually override your actual location with the Change location setting on the left side of the serps page, so you might want to give that a try. I've never succeeded in getting rid of localization effects for any particular location.
I think that the degree which location affects rankings would of course vary by query, but also by where and when you're searching, and it would depend on what Google might be testing at a given time. Google plays with localization a lot in its testing. It's worth noting that in its 2011 Quality Raters Guide (General Guidelines, v3.18), Google devotes twelve full pages to the importance of query location... that's 12 out of 125 pages.
Back in August 2011, we allow a specific search to be discussed here to examine bullet points in SERPs in this thread...
Bullet Points in SERP description, is that new?
I observed in the course of my 4 posts on the thread that the organic results could be shifted quite a bit just by shifting my default location several exits up or down the freeway, to move closer or further to a particular company store we were discussing that showed these these bullet point snippets for a query. Some results disappeared entirely if I moved too far up the freeway, away from the national chain outlet that I knew was a few exits further removed. I hadn't in fact expected these serps to be localized at all.
When I revisited the thread on Nov 15, 2011 and ran the same tests again, I saw that extreme localization for that particular query, at any rate, was apparently no longer a factor. It's possible that Google has pinned down its location based factors for a while, at least for some searches in some locations, and that they're now playing with other possible variables... so you may not observe the shift I saw by using Change location.
"Extreme localization" was brought up by Greg Boser as a strong ranking factor at PubCon Lv 2011, and those observations certainly fit in with what I and several others here have been observing. To some extent, also, I believe localization testing might be what people are experiencing when they report "zombie traffic".
I'd be curious to hear what happens when you systematically run tests with different locations set on the searches you've been observing. You may or may not be able to override very distinct IP and ISP information, though... I'm not sure. IP vs Changeed location might even be another thing that Google is testing.
Perhaps you should also be careful what you eat for dinner before you do a search, as Google might be testing that too. Definitely be careful about where you eat it. ;)