I think that's a fair point. But organizational complexity is compared to other measures, such as the complexity of the software itself, the number of dependencies, etc - i.e. the size of the software itself - and the study found that organizational complexity is still the #1 method.