Usually because the pop-description is, as usual, misleading. People find data covering "several orders of magnitude" and just assume by some cosmic fact that Benford's law should apply, despite the mechanics being their data having no reason to imply that the law holds.