Agreed. The observations in this thread are not all from present time. Some of them are from more than two years ago. It is very possible - even likely - that this thing (whatever it is) has evolved over time. That's why I think drall's recent observation is so very useful. Ir's right now. Maybe Ian Turner will turn up some clues - he's got a present time situation
It was over two years ago that some sites started seeing a "yo-yo" ranking for some keywords, but at that time we didn't think to ask for overall traffic levels. Still, that could have been the earlier days of a throttling mechanism, if total traffic stayed stable even when one important keyword dropped dramatically for part of the day.
It's very possible that this thread may not get to the root of it. We just don't have enough data right now. I can understand why busineses aren't looking for it. If their traffic levels are steady and predictable - happily sustaining their business - then it's on to other tasks.
I started this thread in the hopes that we would find a new audience - others who are seeing something like this but hadn't focused on it so far. I'd start by creating week-by-week graphs of Google referred traffic, not looking at a day-by-day snapshot. If the weekly trend is "supernaturally" flat for an extended period, then it's worth digging further, especially if you track rankings for major keywords and can see that they are not flat.