In this section it is summarized, what kind of research needs exist in the forecasting literature concerning demand forecasting practice. After that, an agenda for answering those needs is presented.
2.8.1 Conclusions from the literature
1) Organizational issues are still relatively neglected in forecasting research.
Even though it has for long been emphasized that organizational issues are critical in implementing efficient forecasting practices, still quite little attention has been paid to organizational issues in the forecasting literature. However, some studies suggest that there are some common problems in implementing efficient forecasting practices.
Therefore, it is possible that organizations could learn from the experiences of other organizations. Models that help pay attention to right things can facilitate decision making processes. To gain deep enough understanding in this area, more studies on implementing and improving forecasting practices are needed, for example case studies.
In the literature it is noticed that it is fundamental to focus the efforts of salespeople to the products and customers where the salespeople’s’ insight on changing demand patterns is truly important. However, there is a lack of descriptions on how this principle can be put in practice.
2) The special characteristics of the industrial context are not well noted in research concerning forecasting practice.
Most of the studies that deal with forecasting practices are conducted with surveys. Only in a few surveys, industrial companies are distinguished from consumer companies.
Most of the surveys provide information on which forecasting methods are used, but not on why or how they are used. Industrial markets have some special characteristics compared to consumer markets. For example, the number of customers is typically lower and the relationship with the customers closer. It can be expected that also the problems in managing the demand forecasting process in the industrial context have some special characteristics that can be studied more deeply.
3) More attention should be paid on the contexts in which forecasting is applied.
The forecasting literature dealing with method selection provides clear solutions for clear situations. For example, if there is sufficient demand history available and no changes are expected, time-series methods can be applied, and if large changes in demand are expected and only qualitative information is available, some judgemental methods can be suggested. In practice, it is possible that the situation is such that there is some historical data available and some contextual information available, but it is not clear how reliable these pieces of information are. Typically, in the industrial context the customer base is heterogeneous, so selecting methods is not that straightforward. Thus, a
possible area for research is finding out how the method selection situations are really like, how well the environment is known, and how the analyzing of the environment can be facilitated.
The literature emphasizes quantitative methods, but qualitative forecasting methods are preferred in real life. The former literature points out that the right choice between qualitative and quantitative methods depends on the circumstances, and often some kind of combination of judgemental and quantitative methods is necessary. More attention could be paid to the different environments where forecasting is applied. It could be examined, how quantitative data can in practice get connected into demand forecasting processes in different contexts.
4) There is room for development in forecasting performance measurement Performance measurement has been noticed to be a critical factor in demand forecasting management. Performance measurement should provide feedback for the forecasters, so that they are able to make the forecasts more efficiently. Efficiency in demand forecasting means that the benefits received from forecasting are greater than the effort put in forecasting. This means that not only forecast accuracy, but also the forecasting efforts need to be considered. Especially in the industrial context, forecast accuracy as a single measure of performance is insufficient.
However, in many practical settings forecasting performance measurement focuses on measuring forecast accuracy. If only forecast accuracy is provided as a feedback, the forecaster becomes penalized for irreducible demand uncertainty. Therefore, there is room for research that aims at developing performance measurement practices that support the improvement of forecasting practices better.
5) More attention could be paid to facilitating the cross-functional dialogue needed in forecasting.
A majority of forecasting literature focuses on forecasting techniques and on the accuracy that is received. However, some research results imply that achieving as accurate forecasts as possible is not the only aim of the forecasting process. Another relevant aim is achieving a cross-organizationally shared view of the demand environment and about the ways to react to it. Therefore, a potential area for forecasting research is facilitating the cross-functional dialogue that is needed in producing the forecasts and in improving the forecasting practices.
2.8.2 Agenda for answering the research needs
In the previous section, some research needs in demand forecasting practice were addressed. In this section it is described how those research needs can be answered. The focal areas of this thesis can be divided to four interlinked areas (see figure 3): 1) defining the operational environment for forecasting, 2) defining the forecasting
methods, 3) defining the organizational responsibilities, and 4) defining the performance measurement process.
1. Defining the operational environment for forecasting:
forecasting needs and availability of information
sources
4. Defining the forecasting performance measurement
process
3. Defining the organizational roles/responsibilities in
forecasting
2. Defining the forecasting methods
Figure 3: Areas of focus in this study
Defining the operational environment for forecasting means exploring what the environments in which forecasting is applied are really like. For an outside facilitator or a consultant aiming at improving the forecasting practices, it is important to gain a view of the environment, as in different environments it is reasonable to focus the forecasting efforts on different issues. For that purpose, approaches for analyzing the demand environment are welcomed. The industrial context has received little attention in the forecasting literature, though forecasting methods are applied also in industrial contexts.
Therefore, it is of value to focus on the industrial context.
Defining the forecasting methods means selecting the forecasting methods and defining the level of detail on which the forecasts are to be produced. The context where forecasting is applied impacts the applicability of the forecasting methods. A relevant
question is what the situations, where forecasting methods need to be compared or selected are really like, and what kind of tools are needed in the selection process.
Defining the organizational responsibilities means defining who does what in the forecasting process. Former research points out that forecasting effort should be focused on the most important customers and products, and cross-functional dialogue in forecasting process is considered important. A relevant question is what the situations which call for focusing resources or improving cross-functional communication are really like, and what kind of tools can be used to support these tasks.
Defining the performance measurement process means selecting the performance measures and linking those measures into the planning processes. Former literature points out that performance measures should support improving the forecasting process.
A relevant question is what kind of performance measures are needed in practice, and how the performance measures can be linked to the planning processes in specific settings.
As a summary, supporting the improvement of forecasting practices requires pointing out the problems that occur in the improvement, and suggesting cures for those problems. Improving forecasting practices can be directed to single phases of the forecasting process, or the process as a whole.
The research questions of this thesis are:
1. What kind of challenges are there in organizing an adequate forecasting process in the industrial context?
2. What kind of tools of analysis can be utilized to support the improvement of the forecasting process?
3 Research strategy
This chapter concerns the methodology and research design of the study. In this chapter, the methodological choices of the study are explained. First, the research paradigm is presented. After that, case study as a method and its different forms are discussed.
Finally, the details of collecting and analyzing the data are presented.