2.2.1 Qualitative Forecasting Methods
The three primary approaches used in qualitative forecasting are the expert opinion approach, the Delphi method, and the market survey approach.
The expert opinion approach is simple and easy to implement. For example, for many of the stand-alone, one-time activities that take place in a project, an opinion based forecast is all that is either necessary or desirable. The opinion of the person who is most knowledgeable in that field is sought. Furthermore, if a project is brand new, the likes of which have never been seen before and for which no historical data is available, then the only recourse for a project manager is to seek the opinion of an expert to get a forecast or an estimate regarding the concerned event or activity.
The disadvantage of relying on the opinion of a single expert is the inherent element of bias. Further, larger issues in the project may arise where an opinion based forecast of a single expert may be not be adequate. This can occur with forecasts involving such things as the timing of the introduction of a new technology into the market place or a change in public behavior as these could have a significant bearing on the decision to start a project or the timing of market entry. When a new product is introduced it can become a guessing game as to how the market will respond and how and when competitors might respond. Answers to questions such as these may require the opinions of several experts, perhaps across a range of subjects, not simply an opinion from those closest to the job. In such cases, the Delphi method may an appropriate forecasting method.
Devised by the Rand Corporation in the U.S., the Delphi technique is a popular method of qualitative forecasting that generates a view of the future by using the knowledge of experts in particular fields. The name derives from the ancient Greek Oracle of Delphi that was supposed to foretell the future. The steps of the Delphi method are as follows:
- Questionnaires are circulated to the team members, who may not be aware of each others' identities, and each is invited to make his own prediction of future progress in a particular field. As far as possible, projections must be quantified and the questions must be framed accordingly: e.g., what proportion of all households do you expect to have a personal computer by the year 2010?
- After the first round of replies, the results are analyzed statistically (giving the distribution of responses) and the results re-circulated; panel members are asked to reconsider their views in the light of the new statistics. If their view lies outside the inter-quartile range, they must either revise their opinion or give their reasons for their extreme view; this will be seen by the other panel members. This process can be repeated for a third or fourth round until a consensus of opinion is obtained.
Results of Delphi studies are given in the form of timescales and probability levels for the feature being forecast. Some large corporations have used the method for assessing long term trends and the development strategies that may be open. Research by the Rand Corporation indicates that with current technologies and trends, the Delphi panel does tend to move towards a consensus view which is generally correct, but there tends to be less accuracy when forecasting new developments. On occasions, no consensus view is obtained after several rounds.
The market survey approach is the third qualitative approach that can be used to generate forecasts of project events. This approach involves surveying past customers or potential customers about any plans they may be considering the future. The project organization's marketing staff is perhaps the ideal source to obtain such information because of their direct contact with customers. In addition, the marketing staff, along with the procurement staff, which is in direct contact with suppliers, can also provide market intelligence reports regarding competitors who are contemplating new projects or new technologies.