To be fair, relying on product teams to shut themselves down, in case they've hit a dead-end, probably wouldn't work.

Actually, why not? It's not like the employees get fired. You generally know when you're working on a crappy product.

Google Reader was a fantastically useful product for me. Not "crappy" at all.

That's the point parent is making: the Reader team probably wouldn't have opted for shutting down, and they would have been right.

The reader team didn't exist (and hadn't for years) at the time the decision was made to shut it down.