Robert_Charlton - 9:17 am on Sep 12, 2012 (gmt 0) [edited by: Robert_Charlton at 9:56 am (utc) on Sep 12, 2012]
If sites are all affected in the same time patterns, that would be VERY significant, as I see it. We have not been able to establish that as a fact so far, even though we've been discussing this topic for 4 years!
That's something I tried to do in my May 2, 2012 post on this thread, referencing both the May 2012 serps thread [webmasterworld.com...] regarding what appeared to be geo-confused zombie traffic in Australia, which I assumed might have been a test... and also referencing my user-engagement calibration phase? thread [webmasterworld.com...] ...about which I noted....
I've since come to think that traffic shaping isn't just about Panda, and have been thinking for a while that it had to do with all Google testing, much of which I've assumed is about localization.
I see comments appearing on the forums almost always after an update and we've entered what I assume is a (new) calibration phase, and they're essentially always in sync with backdraft7's posts.
To toss out some other factors that might be worth comparing while we're discussing getting more specific about some details...
- Do the sites affected have a high product churn rate... ie, a large percentage of ongoing product or listing changes?
- Are there ambiguities in the place names involved... possibilities for geo-confusion (as I noted earlier in this thread, back on May 2?)
About throttling with regard to time of day, I've seen pretty convincing evidence that relates that to things like lunch hours and search activity, and I don't consider it the same thing at all.
[edited by: Robert_Charlton at 9:56 am (utc) on Sep 12, 2012]