That's where the discussion pyramid comes in: Is it real, do we understand the causes, is it important, etc.
More and more lately, I am encouraging teams, as a complementary approach, to consider creating multiple scenarios.
Continue as isEven if the future scenarios are very rough, it at least forces the consideration that there are a range of choices and that the first one that leaps to mind is not necessarily the best.
Another technique, which I have been playing with but not yet deployed, is to get the team to agree on a measured description of the current state. Once they are in agreement that this is the size, shape, and nature of the current state with which we are dissatisfied, then get them to use their speculative imagination of all the possible circumstances that could but (not necessarily does) explain that particular pattern of data. More than a simple root cause analysis which is anchored in what is known. Rather, it is an exercise in realistic fiction - not what you think did cause it but what hypothetically could cause it.
The intent is to get behind the headline assumptions and unearth deeper or alternative causal factors which we might be, owing to our priors, disposed to ignore. It is the unseen that usually subverts realized outcomes.