I agree and I don't think it is specific to the world of entrepreneurs and/or social science experiments. If your target variable depends on a large number of correlated variables you are very unlikely to formulate the correct hypothesis by accident. This is why you need intuition or good judgement in science, just as much as in every thing else.