Identify which forecasting technique or multiple techniques should be used in the future for the company’s strategy. Are there other techniques available that are not listed above?
Explain the technique you identify, and give an example of how it is used in the manufacturing, retail, and health care industries.
Detail if 1 of the 4 techniques listed above should NOT be used and why.
Evaluate the significance of forecasting error for the technique or techniques you have selected. What is the impact of error on your chosen technique?
Forecasting is of great importance to a business. It involves predicting future trends in business in a timely manner that allows the business to adjust accordingly. Poor forecasting is a major contributor to high costs in companies. Forecasting relies on good judgment, analysis of prevailing economic conditions, and intuition in order to make accurate forecasts. Most analysts find it difficult when it comes to developing actual forecast numbers such as projected sales that the company can accurately rely on to make production decisions. Forecasting can be done qualitatively or quantitatively depending on the availability of historical data. Qualitative method is applied when there is lack of historical data, while the latter is used when past information is available and is quantifiable.
The company has historically employed time series method in forecasting. Time series analysis involves making observations relating to a particular variable and measured successively in time (Anderson et al., 2014). The measurements may be taken on various basis such as hourly, daily, monthly, or any other time duration. The rate variable in a time series analysis is recorded at regular time periods. As the company aims at expanding, there is need for a new forecasting method that will not rely on historical data to make predictions. The new forecasting method should take into consideration the planned future growth and hence make predictions by drawing data from future variables. Since the company is expanding operations, the forecasting techniques that would best meet its current as well as future needs are simulation and causal techniques. Another common method used in forecasting is the Delphi method. This method takes a qualitative approach during analysis. Two sets of questionnaires are given to a panel of experts who are physically separated. The results are then reevaluated.
The causal technique involves establishing whether there is a causal relation among various variables that may affect demand of the company’s products (Naidu, 2013). If the variable is identified in advance, projections can be fairly made. The assumption is that the variable under forecast has a causal relationship with other variables. For instance, the number of building under construction in a particular area affects the demand of different materials used such as wall papers, cement, glass, and others. Causal forecasting can be carried out in three different ways: regression analysis, use of econometric models, and use of input-output models. In regression analysis, a mathematical equation is used to relate the dependent and independent variable(s) that influence it. Econometric models employs interdependent regression equations that tie to economic activities in various sectors of the economy. Input-output models are used explain flows of goods in different sectors of the economy.
Causal relationship is used in the manufacturing sector to predict sales. For example, there is a positive correlation between advertising and increment in sales. Manufacturers can use regression analysis to show the relationship between sales and the amount of advertising expenditure that the company plans to use (Naidu, 2013). Also, a company that produces cement can look at the number of building permits issued to make future sales predictions. In the retail industry, causal relationship is used to make sales forecasts. Retailers use causal relationship method to forecast future demand. For instance, discount promotions during holidays greatly impacts demand of some goods. Retailers must factor this while making orders or in shipping goods. Causal relationship method is also used in the health care industry to establish causality or relatedness of a particular drug and possible adverse effects on patients. The healthcare industry is flooded with new drugs which has various adverse effects on users. With an increase in adverse drug reactions among individuals, there is a greater need for causality analysis.
The company should not use qualitative method in forecasting. Qualitative method is often applied when historical data is unavailable or inapplicable (Montgomery, Jennings, & Kulahci, 2011). For instance, during the introduction of a new product since there is no historical data available. The method involves use of judgment by experts and is subjective in nature. Even though some data analysis may be performed, the key for making the forecast lies in subjective judgment. Qualitative forecasting has inherent drawbacks that make it inappropriate to use in the case. First, it is difficult to make accurate forecasts since those involved in forecasting rely mainly on qualitative data. The reliability of this type of data is low. It can be difficult especially for forecasters to process large or complex information. This reduces the forecasters’ ability to give accurate forecasts. Political factors within organizations can also affect the decision making process by forecasters. This is because the method is based on making subjective judgments.
The quality of forecasts given is highly affected by forecasters’ personal attitudes and dispositions. For instance, the forecaster may be overconfident which may negatively affect the reliability of the forecast. There are also issues with inconsistency in judgment of the forecaster due to personal factors such as moods and emotions. Qualitative methods are highly affected by anchoring. This is a situation where the forecaster’s judgment is affected by forecasts conducted in earlier periods, for instance those conducted through quantitative methods. The forecasters’ tendencies to infer patterns in data may also have an impact on the quality of forecast made. The forecaster may likely infer relationships and in the process make the wrong sales forecasts. A forecaster may also be tempted to justify what he/she believes is true rather than understand the facts of a particular scenario. This may lead to a biased judgment. Lastly, qualitative forecasts may be time consuming and thus expensive.
Forecast errors represent the difference between the projected value and the actual value that occurs. In all forecasts, there must be some percentage of error. There are different types of error depending on the source. The most common errors are bias errors and random errors (Jain & Malehorn, 2005). Bias errors arise from mistakes during forecasting, for instance, using the incorrect variables. Random errors may be difficult to detect. Forecasting error has a significant impact on the causal relationship forecasting method. Bias errors result in either over-forecasting or under-forecasting. Accuracy error is concerned with the relatedness of the error to what was supposed to measure.
Forecasting errors may either be positive or negative depending on the nature of the error (Jain & Malehorn, 2005). For instance, a forecast error may cancel out the effect of another error that was previously made. Researchers often use the mean squares error method to eliminate the chances of there being errors that cancel out. On the flip side, forecasting errors can lead too inaccurate forecast results. Such errors may lead affect the quality of the decision making process. The management will likely make the wrong decisions if they are based on inaccurate figures.
Anderson, D. R., Sweeney, D. J., Williams, T. A., et al., (2014). Statistics for business and economics. New York, NY: Cengage Learning.
Naidu, R. P. (2013). Causality assessment: A brief insight into practices in pharmaceutical industry. Perspectives in Clinical Research, 4(4), 233–236. http://doi.org/10.4103/2229-3485.120173
Montgomery, D. C., Jennings, C. L., & Kulahci, M. (2011). Introduction to Time Series Analysis and Forecasting. Hoboken: John Wiley & Sons.
Jain, C. L., & Malehorn, J. (2005). Practical guide to business forecasting. Flushing, N.Y: Graceway Pub. Co.