Tuesday, September 25, 2007

Qualitative Methods of Forecasting

The method of decision making is not always based on quantitative data. Forecasting in the real world can often be affected by events that are not reflected in the data. Changes in government, weather problems, wars, and other things may require company decisions without data.

Consider the recent examples concerning toy recalls at Mattel due to defects (read about it HERE). Faulty toy designs and lead tainting in the paint have made crucial decisions in production necessary. Mattel had to choose whether to halt production and recall the toys based on current data, possible forthcoming data, but primarily based on a qualitative feeling of market sentiment. The expected harm to future Mattel sales due to this event was probably the driving factor in the recall. There was very little past data to mine to achieve the qualitative forecast that sales would decrease if unsafe toys were allowed to stay on the market.

There are 4 Qualitative Methods of Forecasting:
  • Jury of Executive Opinion
  • Delphi
  • Sales Force Composite
  • Market Survey


  • Jury of Executive Opinion
    This method of forecasting takes all of the opinions of the highest level executives in the firm that are related to the decision and puts them to a vote. This may be by consensus, by direct vote, or by final executive decision. It depends partially on the culture of the company. These executives may be guided to a decision by supporting data, but often they have no supporting data; only experience and intuition to guide them.

    Delphi Method
    Originally developed by the RAND Corporation back in the 1960's, many other companies now employ this method of forecasting. The strength of this method is that it employs specific experts in various related fields. These delphi panels consist of experts and a facilitator. Taking a collection of expert opinions, a report is generated that forecasts based on the consensus. The Delphi method is a pretty good method due to a few factors:
    1) Confidentiality - since surveys and votes are taken anonymously, there is little likelihood of reprisal for negative forecasts.
    2) Less vulnerable to manipulation due to the variety of interests involved.
    3) Opinions are based on a broad variety of shared ideas that may lead to innovation in forecasts.Forecastingprinciples.com has a good page with more information about this.

    Sales Force Composite
    Taking a bottom-up approach to forecasting, the sales representatives for the local regions and districts create predictions of their own geographic areas. These local and regional forecasts are the combined into a national forecast. This could be an advantage to a large company in several geographic areas. Forecasts will be based on the opinions of people that actually make the sales.

    Market Survey
    Considered one of the least effective forms of forecasting. Predictions of sales are made based on customer surveys. This is not always effective. First, one needs to consider the errors involved in making a customer survey. Sampling errors, non-coverage errors, non-response, and measurement errors can all cloud the accuracy in the survey making process. These results may be made even less because some customers often say that they will do one thing but choose differently when they actually need to spend their money.

    So, as you can see, there are needs for qualitative forecasts. The methods used to do the forecasting have varying degrees of accuracy and success. In my next several blog entries, I intend to tackle the subject of quantitative methods in forecasting. Until next time, thanks for visiting.
    -------Sincerely, Trevor Stasik.



    To return to initial post about forecasting, click HERE.
    To view the next post in the Forecasting series, click HERE.



    View Trevor Stasik's profile on LinkedIn

    No comments: