I've used that concept in a RPG system I was designing many years ago. But I wouldn't apply it to levels above individual skills - success or failure of a large group effort does not imply anything about the individual successes and failures.
To use a modern example, a User Interface designer could design a poor to use interface, a developer could implement it brilliantly, a QA person could find bugs in the backend (success) but not report the poor interface (failure) and the salesperson could claim it does things outside the amount of features it has which makes more sales (success) but less repeat business (failure). The success of the project as a whole doesn't address the lessons learned by the database developer who first put together something that worked at first glance but couldn't handle some of the needs and ended up getting redesigned in a much more flexible way - that failure advanced them, gave them XP.
Even an encounter is too large for the adage: a defender dropping to a lucky crit isn't really an indication that they had a learning experience even though they "failed". And if a wizard did great and others didn't hold up their part and a scene was failed, do you reward the others and not the wizard?
Just to change topics some, the idea of failure leading to growth is a good one for slowing advancement when you are already really good since you need to find challenges that are meaningful and hard enough to fail. However, many (most? all?) RPG gamers see XP, at least partially, as a reward. Withholding it when they do well may not sit comfortably with many.