Based on the document, PayTran implemented a series of fraud detection process changes for one of their clients, IWantIt, to help reduce the level of fraudulent transactions. After these changes, they collected data for a month to evaluate the performance of the new process. Given this context, thinking about whether the fraudulent transactions have decreased, can you be 100% sure based solely on one month of data? 

What factors would you consider to confirm or deny this decrease without looking into the specific numbers? Think about the complexity of experimentation and the difficulty in being certain about the effects of changes. 

Consider the limitations of using a short data period for evaluation and the potential influence of external factors on the observed outcomes.