It is important to investigate when there is a large discrepancy between the predictive analysis done beforehand with the actual results of sales projections before taking action. There can be a number of reasons for the discrepancy such as changes in market conditions, unexpected events, or errors in the data used to create the regression model (Richardson, V 2023). Predictive analytics requires large amounts of data to produce useful results, and because machine learning algorithms are utilized there are times that external variables cannot be accounted for (University of New South Wales 2020). To explain the discrepancy to the audit manager, I would recommend conducting a thorough analysis of the factors that may have influenced the actual sales results, such as changes in market conditions, competitor activity, or other external factors. It may also be necessary to review the data used to create the regression model, to identify any errors or inconsistencies that may have impacted the accuracy of the predictions.

To minimize large discrepancies in the future, I would recommend a few different strategies. First, regularly updating the data used to create the regression model is important to ensure that it is reflecting current market conditions and customer behavior. Then using multiple regression models to account for different scenarios and market conditions can help ensure that there are no errors. Doing follow-ups to regularly review and refine the regression model based on actual sales data and communicating with the sales teams can help ensure accuracy. Scenario planning can also with understanding best case and worst case scenarios that can possibly happen (Wang, T 2016). This can allow the business to prepare for a range of outcomes and prepare for those occurrences.