In many domains, such as disbursement data, sales, lengths etc. the first digit of a number follows a precise distribution. In one number out of 3, for instance, the first digit is a 1. If a data set does not follow this distribution, it might be an indication of fraud. Tax authorities around the globe use it, with other tools, to spot wrongdoers. Please note that Benford's law does not apply to constrained data (height of human beings, invoice numbers, etc.).

Read more at Wikipedia

Find the bad guys with Benford's Law

Do your data follow Benford's law? If they don't, it might be that someone meddled with them!