Qualitative Demand Forecasting, Causal and Time Series Methods - The Thesis

Header Ads

Qualitative Demand Forecasting, Causal and Time Series Methods


Picture of Pie Charts, Barcharts and Line Graphs

Introduction
Demand forecasting is essential to every business since demand is a significant driver in every industry. Without the need for an organization’s goods and services, there can be no business. Companies seek to maintain a delicate balance between supply and demand. For this to be achievable, there need to be ways and means by which future demand can be predicted; hence, the development of demand forecasting methods. A forecasting method is a numerical modus operandi for producing a forecast. The horizon for demand forecasting can range from years (long-range), months (medium-range), and weeks (short-range).

Contents
1.    Importance of Demand Forecasting

2.    Quantitative and Qualitative Demand Forecasting Methods
         2.1 Qualitative Demand Forecasting Methods
        2.2 Causal Demand Forecasting Method
        2.3 Time Series Demand Forecasting Method

1.    Importance of Demand Forecasting

Demand forecasting enables planning of new facilities. It takes quite a long time to design and build a new factory or design and execute a new production process. Against this backdrop, it is of import that operation managers have enough lead time to build factories and install procedures to produce the goods and services as and when needed. Demand forecasting makes this possible.

In Need of Writing Services? Contact us for quality work!

Every business experiences fluctuations in demand for their goods and services and this demand may vary from month to month. Businesses seek to do their very best to stay one step ahead by forecasting future demand for their goods and services to enable product planning; this move is to help businesses stay competitive in the marketplace and avoid the risk of losing their customers to their competition. Aside from the risk of losing to the competition, changing the capacities of production processes to reflect variable monthly demands can take several months. Demand forecasting enables the operations managers to gain the necessary lead time to facilitate the change in the production process.    

Additionally, demand forecasting also makes possible effective workforce scheduling. The fact is demand for goods and services may also vary from week to week necessitating the scaling up or down of labor force (using reassignment, overtime, layoffs or hiring) to satisfy demands. Data produced from demand forecasting methods equip the operation managers to have the lead time to effect the necessary changes in the workforce to meet the weekly requirement.

2.    Quantitative and Qualitative Demand Forecasting Methods

Over the years, some techniques for demand forecasting have been developed. They range from naïve techniques, opinion sampling, qualitative methods to quantitative methods. Time series and causal methods both constitute quantitative methods. Other quantitative forecasting techniques are trend analysis, seasonal adjustment, decomposition, graphical methods, econometric modeling, and life cycle modeling. While qualitative forecasting methods rely on the judgment of experts, quantitative methods rely on mathematical models. Meanwhile, other qualitative forecasting techniques are market research, focus groups, historical analogy, and panel consensus.

2.1 Qualitative Demand Forecasting Methods

Here are some applications of qualitative methods: the Jury of Executive Opinion Method, Executive Opinion, Expert Opinion, Delphi method, sales force survey and consumer survey. 

In the Jury of Executive Opinion Method, executives are each asked to give their expert opinion on expected market demand during a specified time frame. These individual estimates are then aggregated into a single assessment. The advantage of this kind of qualitative method is that it allows for a wide range of factors to be considered. Moreover, managers find this method most `appealing. Notwithstanding, there may be deep-seated biases which may, in the long run, affect the accuracy of forecasted estimates. These biases may not be evident at first glance. According to Armstrong and Green (2012), experts are often biased when making forecasts: “Salespeople may try to forecast on the low side if their forecasts are used to set quotas. Marketing executives may forecast high, believing that this will gain approval for a project or motivate the sales force.”

Under the Expert Opinion method, stakeholders, be it dealers, distributors, retailers, wholesalers, suppliers and consultants, engaged in different aspects of the organization’s business are ‘bunched up”. The consensus is attained by each uniquely bunched up group (e.g., retailers), after which a final overall average estimate is determined from expert opinion forecast of each of the unique groups.

Under the Delphi method, a mail survey is used to extract the expert opinions of a group of experts, eliminating the demerits of traditional group meetings (Armstrong and Green 2012). A summary of the responses of the experts is done without disclosure of the identity of the experts and then mailed again to the experts along with a questionnaire “engineered” to explore the reasoning behind extreme opinions proffer in the first round. Until a reasonable agreement emerges among the experts, the process may be extended for one or more series. According to Rowe and Bright (2001), forecasts produced from Delphi were more accurate than those provided from traditional meetings in five studies. Armstrong and Green (2012) also contend that Delphi is expected to be most successful in instances where relevant knowledge is dispersed among experts.

2.2 Causal Demand Forecasting Method

This method is used to produce a demand forecasting models based on the existence of a substantial cause and effect relationship between explanatory variables (i.e., independent variables) and the demand variable itself  (i.e., dependent variable) (Armstrong and Green 2012). Depending on the depth of knowledge of the variables affecting the situation of interest (i.e., demand) and availability of data, causal methods may have the following branches: regression analysis, the index method, and segmentation (Armstrong and Green 2012).

Causal demand forecasting methods, though it may be entirely accurate in forecasting demand, has some limitations. The thing is that causal processes can be regarded as the most useful only if: 1. The causal or explanatory variable shows a significant degree of variability over the forecast horizon; 2. The directions of the relationships are established; and, 3. There is prior knowledge of the variables (Armstrong and Green 2012).

However, causal demand forecasting methods are not without their advantages. Causal method, in the form of regression analysis, enables the evaluation of the relationship between one or more independent variables and a dependent variable (i.e., demand). Regression analysis is often applied to historical data. In the form of an index method, causal method is convenient for instances where data on the variable (i.e., demand) to be forecast is scanty and where several independent variables are essential as well as existence of prior knowledge of their respective effects on the dependent variable (i.e. demand) (Armstrong & Green 2012).

2.3 Time Series Demand Forecasting Method

For this method to work, the variable to undergo forecast must have consistently shown specific unique patterns in a past time horizon and that this identified pattern will be sustained into the future (Anonymous 2011; Brockwell & Davis 1986). These patterns can either be cyclical, seasonal or periodic (Taylor 2008).

Time series collected at regular intervals are then used to produce models. The time series method is most applicable when there is a lack of a distinct upward or downward pattern in the historical data being explored, showing an absence of a linear relationship between demand and time (Armstrong and Green 2012).

Moving Average, Exponential smoothing, and Trend projections are all offshoots of the Time Series method of demand forecasting.

All in all, demand forecasts are not immune to error and uncertainty which may originate from there main sources: 1. data about past and present market; 2. methods of forecasting, and 3. environmental change. As such, it is critical that careful deliberation goes into the selection of the appropriate demand forecasting method.

3. Advantages of Causal Forecasting

  • Possess explanatory powers.
  • Allows for the execution of if forecasting by exploring the interaction among demand variables.
  • Acts as a valuable tool for planners, marketers, and strategists.
  • Allows the forecasting of the effect of policy changes such as price changes and promotions.
Key Terms:

No comments:

Powered by Blogger.