(Thirty-eighth in a series)
Last week we discussed the role of expert judgment in making forecasts. When quantitative data are not available, or when we are trying to predict a major structural shift in the future, we often rely on those people who are well-versed and knowledgeable in the field for which we seek forecasts. The Delphi Method is one way to do this.
Developed at the start of the Cold War by the Rand Corporation, the Delphi Method has its groundings in technological forecasting, as it was designed to forecast the impact of technology on warfare. The name “Delphi” comes from the Oracle of Delphi, which in Greek Mythology foretold the future. Quantitative models are often of limited use when trying to predict far into the future. Environmental patterns, largely driven by technology changes, can be altered dramatically over long periods of time. When projecting far into the future, we want to know how probable, frequent, or intense these future forecasts are or will be. This is where Delphi comes in.
The Delphi Method is a structured, interactive, iterative communication technique joining together experts to share their opinions on the future. Unlike the Jury of Executive Opinion, which we discussed in last week’s post, this panel of experts does not meet face-to-face. This ensures that experts’ opinions are not influenced by those of other panel members. The number of experts on the panel is large, and many of them may differ greatly in their areas of expertise.
Panel members are given questionnaires and asking them a series of “what,” “if,” “what if,” or “when” questions about the future. They may even be presented with scenarios and asked to predict the probability of such a scenario occurring and when it may occur. Differences in experiences, information availability, and interpretation methods between panel members will ensure a wide diversity of views. In order to move panelists to consensus, their opinions are summarized and shared (anonymously) with the other panel members, and the panelists are encouraged to adjust their predictions based on these viewpoints. When certain panel members hold views substantially different from the group median, they are asked to provide written justification, so that the strength of their opinions can be determined. After a few iterations, the group tends to move toward a consensus forecast.
The Delphi Method is not without its drawbacks. While the absence of face-to-face meetings eliminates biased viewpoints brought on by authority, seniority, and articulation, it also greatly reduces – if not also eliminates – immediate access to the knowledge of others. Hence, panelists provide their views in isolation and, based on their experiences, may not consider certain facts in their assessments. Moreover, Delphi techniques can be expensive and time consuming, as experts’ time is at a premium, and searching for them can be intense. In addition, because the Delphi Method is used to predict several years into the future, a lot of time must be allowed to elapse before one can determine whether the method was appropriate for the task on which it was used. Finally, just because the iterative process moves experts towards a group median, it’s less clear that the process pulls the group towards the true future outcome.
Next Forecast Friday Topic: Other Judgmental Forecasting Methods
In next week’s Forecast Friday post, we will be discussing a few other judgmental forecasting approaches that are used when quantitative data is not available. The week after that, we will discuss the various judgmental biases that exist in forecasting. These next two posts will round out our discussions judgmental methods, after which we will move into our final segment of the series “Combining and Evaluating Forecasts.”
Follow us on Facebook and Twitter!
For the latest insights on marketing research, predictive modeling, and forecasting, be sure to check out Analysights on Facebook and Twitter! “Like-ing” us on Facebook and following us on Twitter will allow you to stay informed of each new Insight Central post published, new information about analytics, discussions Analysights will be hosting, and other opportunities for feedback. So check us out on Facebook and Twitter!