It seems that at their scale, Google doesn't bother to (is unable to?) evolve individual products, rather they pivot by launching and killing entire projects.
The same thing happened with Google's 3D world product, Lively, and the following comparison of Lively with IMVU by Eric Ries is pertinent:
"By the time they [Google] actually managed to launch something, they were building the product that we had already discovered was not the right product 100 iterations ago.
"But they launched it with such high fanfare and expectations; they were so convinced they got it right, that it was actually the mismatch between the expectations that Google--one of the most public companies in the world--put behind this product, and then its pathetic results...
"The product wasn't a bad product and it's not like nobody wanted to use it (of course they wanted to use it for stuff that embarrassed Google) but it was the mismatch that really caused them to have to pull it; it was very embarrassing for them...
"Even after they launched they really didn't give it enough time to learn and iterate because they had set the wrong expectations, a classic big-company [inaudible]."
This is really a key point. Outside of a few wave template buttons, has any of the core interface problems with the Wave client been addressed since the initial semi-private-invite-only release? Wave seems fundamentally the same to me as a user as it did a year ago when I first saw it. Yet users in general have complained about the same problems the entire time, none of which appear to have been addressed.
The same thing happened with Google's 3D world product, Lively, and the following comparison of Lively with IMVU by Eric Ries is pertinent:
"By the time they [Google] actually managed to launch something, they were building the product that we had already discovered was not the right product 100 iterations ago.
"But they launched it with such high fanfare and expectations; they were so convinced they got it right, that it was actually the mismatch between the expectations that Google--one of the most public companies in the world--put behind this product, and then its pathetic results...
"The product wasn't a bad product and it's not like nobody wanted to use it (of course they wanted to use it for stuff that embarrassed Google) but it was the mismatch that really caused them to have to pull it; it was very embarrassing for them...
"Even after they launched they really didn't give it enough time to learn and iterate because they had set the wrong expectations, a classic big-company [inaudible]."
[http://venturehacks.com/articles/lean-startup -- starting at 6:35 in the audio]