To decide is this is a success or a failure one need to determine exactly where was the error. If this new knowledge contributes to the design of better systems then this is a success, if the actual error provides us with nothing to learn then this is a very big failure. So there is an amortized cost of success/failure in long term enterprises.