reply agree or disagree
This week we were asked to discuss the problem section of the final paper and give as well as receive feedback. My problem section is not laid out into one individual section, but rather in iterations throughout my paper.
One problem that businesses encounter when they do not adequately forecast sales, is not providing enough inventory to meet the customer’s demand. This can be problematic, especially during the holiday season when sales increase. Businesses need to provide enough inventory for customers because if they do not, it could create animosity and prevent them from returning. If a customer is not satisfied, they may not return for a long period. “Time absence plays an important role in relationship revival,” between a company and their customers which can affect their general willingness to return (Pick et al., 2016).
Another problem that arose was the finding within the analytics portion of the paper. There were three different methods of forecasting used when analyzing to determine the results and find which approach was best for this paper. The three methods were regression, moving average, and exponential. Regression is used for linear trends, therefore, if the data does not present a linear relationship, the regression method would not be the best fit (EMC Education Services, 2015). This was proven to not be a reliable method for the data that was collected. The moving average was conducted in 2-month intervals as the highest earnings were in December and January and I wanted the predictions to incorporate these. However, since the average was on a downward trend, it predicted downward and did not account for higher sales in those months. This proved to also not be the most effective. Lastly, the exponential smoothing method was used. However, due to the extremes in the dataset, there was no known trend. Therefore, all 3 methods that were selected were not viable which caused issues with predicting sales.
Many outside factors were identified in the theorised portion of the project which can account for some of these erratic numbers. However, 4-years’ worth of data in this particular circumstance was not enough to determine future sales.
