Question 1a: Define a problem statement which reflects the challenge facing Mark as he planned for the opening of the new center
Managing a successful call center requires careful planning and implementation of techniques that guarantees productivity. However, decentralized call centers poses management challenge especially in case where there is the existence of issues resulting from poor implementation of company’s policies and procedures. In addition, planning for the future is difficult for decentralized call center locations because the agents may not implement the set strategies simultaneously, thus resulting to variation in the actual calls for each call center. This calls or the need to open a centralized call center where the human resource manager is able to monitor, plan and correct forecasting methods as a way of improving performance and customer service. In addition, the option of centralizing the call centers aims at improving employee performance through training them on how to handle employees’ calls while implementing the organizational policies.
Question 1b: Why was Mark’s initial forecast of call volume so far off? What could have been the reasons for this?
Marks forecasting was far off because the method that he used to determine the forecasts lacked scientific evidence to reduce the total relative error. The judgmental forecasting using historical analogy of only one decentralized call center that was used mainly relied on the previous reported average calls made by the employees in each call center did not represent all the other br. However, he did not consider how certain changes in management strategy could affect the number of calls to be made in the future. The estimation method also
Question 1c: What could Mark have done differently to improve his initial forecast?
Other than using personal judgment based on the Mark could have used a second forecasting method to calculate the forecasts and then compare the two results, thus reducing the difference. Secondly, comparing past performance of more than one call center and then finding the average score could have helped in reducing the gap between his forecasts and the actual value.
Question 2a: Describe the details of the Last Value method used by Harry and explain its accuracy (MAD value) in comparison with the accuracy of the other methods.
The Last Value method produces forecasts by observing the value of a particular time series to predict the future call volume. This method uses inconsistence variance, thus resulting to a higher chances of getting huge relative error. As seen in Harry’s calculations, time series is more inaccurate as it has recorded the highest mean absolute deviation as compared to other methods.
Question 2b: Describe the details of the Averaging method used by Harry and explain its accuracy (MAD value) in comparison with the accuracy of the other methods.
This method relied on the average value of the past call volumes recorded in a series of time. The mean absolute deviation of this method is the highest indicating that the accuracy of computing forecasts using averaging method is lowest and the forecasts may be far off from the actual value of the call volume.
Question 2c: Describe the details of the Moving Average (5 days) method used by Harry and explain its accuracy (MAD value) in comparison with the accuracy of the other methods.
This method produced more accurate forecasts as it relied on averaging the past five simultaneous number of calls to predict the possible volume in the immediate days. Unlike averaging method that relies on past records, this method is able to relate immediate factors that may influence the expected call volumes. The accuracy of this methods’ Mean absolute deviation which is the lowest compared to all the other methods.
Question 2d: Describe the details of the Exponential Smoothing (alpha = 0.1) method used by Harry and explain its accuracy (MAD value) in comparison with the accuracy of the other methods.
Exponential smoothing involves giving the weight of the last value to predict the expected forecast while using a smoothing constant. In this case, use of alpha 0.1 reduces the variance between the forecasted value and the actual value. The mean absolute deviation of 274 is relatively lower as compared to that of averaging, last value and smoothing constant (alpha 0.5), thus making this method relatively accurate.
Question 2e: Describe the details of the Exponential Smoothing (alpha = 0.5) method used by Harry and explain its accuracy (MAD value) in comparison with the accuracy of the other methods.
Using the smoothing constant of 0.5 increases the variability of the forecasted value, thus adding more weight than when using a smaller constant. As a result, the mean absolute deviation is relatively higher than when using alpha 0.1. Thus, higher smoothing constant increases the absolute error, thus reducing the accuracy of the forecasts.
3a. Call Volume Forecast for July 2015 (Exponential Smoothing, alpha=0.5): 31,859
3b. Call Volume Forecast for July 2015 (Causal Forecasting based on head count): 33,579
Average Estimation Error for Causal Forecasting model based on headcount: 1677
Explanation of the difference in values:
The difference of the two values results from the variation in the figures used. The latter method involved the use of dependent variable of determining influence of the employees’ head count to determine the future call volume while smoothing model relies only on the recent past call volumes to determine the future volumes respectively.
3d. Personal forecast 28,329
This forecast is achieved by using smoothing model with alpha 0.1 as the variance. The results have also assumed that the last forecasted call volume was 24000 and using the exponential smoothing method similar to 3c above, it is possible to calculate the expected forecast for the month of July. This based on the fact that the lower smoothing variance of 0.1 reduces the total mean absolute deviation, between the forecasted and actual call volumes, thus increasing the accuracy of the results. Despite the results being lower than the previous month’s call volume, the exponential smoothing method using the smoothing alpha 0.1also reduces the risks of over forecasting the results.
References
Armstrong, J. E. (ed.) (200). Handbook of Forecasting Principles. Boston, Kluwer Academic Publishers.
Hiller, F. S. & Hillier, M. S. (2014). Introduction to Management Science: A Modeling and Case Studies Approach with Spreadsheets (5th ed.). NY: McGraw-Hill/Irwin.
Marshall, K. T., and R. M. Oliver (1995). Decision Making and Forecasting. NY: McGraw-Hill.
