you may not have the required data to evaluate all the red flags). For each red flag, perform your analysis at various levels of analysis. For example, start at the aggregate level and look at trends over time, then do the same at the site level, and then at the employee/customer/vendor/item level. If you find something that looks odd in one of these analyses, make sure you explore this finding further. For example, if sales appear to be abnormal for a specific site in a specific year, then analyze this subset of the data further (e g examine if the abnormal sales is associated with specific employees