Most failure gets treated the same. It shouldn't.
Operational failure is failing at the known. Missing a deadline, showing up late, dropping a ball you've held before. You knew what to do and didn't do it.
Experimental failure is failing at the unknown. Trying something untested in an environment you can't fully model yet, and it doesn't work.
Be hard on operational failure. It's a broken promise.
Be easy on experimental failure. It's the price of learning.