So, we had Amazon come for an info session recently, and I thought I would make a note of something they said before I forgot it. It's not on Wikipedia, but an Info Session is hardly citeable, and I thought the information should be recorded.
Anyway, they said Oprah is the hardest thing to predict for them. Christmas, etc. they have a lot of past sales data that gives them good ideas of how they're going to act.
One example they specifically referenced was the Kindle. Apparently after the Kindle got a huge recommendation on her show, they saw a big spike - but were mostly ready for that. So, they were all happy that they'd survived and predicted it so they could handle it. Then, it turned out that that spike was from one small market that got the show live or early or something. Apparently when the show hit the rest of the nation, they were absolutely shocked at the increases, sold out for a long time, orders of magnitude low in their predictions, etc.
Actually, google / the internet seems to bear out that Oprah's recommendation was indeed a very important moment in the Kindle's history.
They had to keep saying they couldn't give exact numbers, which seems to be a definite trend, even where it seems meaningless. My friends who worked at Google act similarly - I wonder if that's like the new thing to do in big tech company corporate policy.