A lot of decisions at my company are based on observed correlations. “We did this and that happened. That was good so let’s do more of this.”
That reasoning only works when all other variables are held constant, which is inevitably almost never the case.
Correlation does not imply causation. Consistent progress on the effect can be made only by first finding out the right cause and then focusing effort on changing it in the desired direction.
It’s the same thing with lifestyle. We feel bad or fail and start to blame it on obvious culprits, often the same things we’ve blamed in the past. When you feel bad or fail, consider all variables and do control tests for the one that’s actually predictive. This is the only way to get better.