Qualitative Forecasting

1
Qualitative Forecasting

• Also known as Technological or Judgmental
  Forecasting
• Use
• Long Term
• No history
• Historical patterns are not expected to apply
• Several methods
• Subjective
• Exploratory
• S-Growth Curves

2
Subjective Methods

• Jury of Executive Opinion
• Face-to-face discussion
• Experts from diverse areas
• Jury interacts to develop a forecast

• Advantages
• Simple
• Incorporates experience of several experts

• Disadvantages
• Bias from more outspoken people or bosses
• No standard method to converge on the forecast

3
Subjective Methods

• Sales Force Composite
• Tally sales force for forecast and then sum
  responses

• Advantages
• Quick and easy method
• Have disaggregated data

• Disadvantages
• Optimistic bias
• Lacking big picture judgement

• Market Research /Surveys
• Systematic, formal estimates based on
  statistical surveys
• Particularly useful in estimating new product
  demand
• Costly method

4
Exploratory Methods

• Look at alternative futures (to better plan
  through anticipation and influence)
• Scenario Analysis
• Especially applicable to strategic planning
• Identify interactions among possible events
• Process
• Experts are presented purpose and system being
  modeled
• Experts develop alternatives technologies
• Timing and interactions are explored for
  alternatives
• Write scenarios
• Scenarios circulated for input

5
Exploratory Methods

• Scenario Analysis
• Advantages
• Simplifies complex interactions/systems
• Considers many combinations at once
• Useful when technology change expected
• May improve long-range planning
• Disadvantages
• Evidence hasnt shown improved management
  performance in dealing with changing climates
  when method is used

6
Exploratory Methods

• Delphi Method
• Forecast time and probability of future event
• Process
• Diverse experts are polled for estimates
• Responses are summarized and sent back for
  another round of estimates
• Process repeats until a group opinion is believed
  to have been reached
• If significant differences exist these are
  studied too
• Unbiased method of getting expert opinions and
  idea interchange and appears to improve estimates
  of when event will occur

7
Exploratory Methods

• Cross Impact Analysis
• Defines dependence of a forecast on other
  forecasts
• Often used together with Delphi or Scenario
• Questions are such If A then what is chance of
  B? Get the conditional probabilities
• Experts are forced to consider interactions
• Disadvantage is that future casual interactions
  are hard to estimate
• La Prospective
• French designed process where individuals and
  organizational activities are emphasized in a
  cross analysis
• The assumption is the the sponsor will act to
  influence the sequence to improve the possibility
  of positive events

8
Exploratory Methods

• Analogy Methods
• Compare forecasted variable with similar case
  where the history /outcome are known
• Ex model intro of air travel by what happened
  with rail
• Trend Analysis
• Popular method that works OK for linear
  relationships
• Often applied inappropriately
• Nominal Group Process
• Like Delphi but after initial poll, experts get
  together for discussion. Discussion is a possible
  source of bias
• Case Studies
• Generalizing a case study to another situation

9
Growth or S Curves

• Technology life cycles follow these curves
• Slow initial growth
• Rapid increase
• Slowing maturity
• Final decline
• Expert opinion (subjective) required to estimate
  key components of curves - max height and proper
  curve (shape)

10
Growth or S Curves

• Curve explained by three parameters
• a - location, b - slope and L - max height
• Two most common curves are Logistics Gompertz
• Gompertz - not symmetric, reaches inflection
  sooner than Logistic curve
• A little data can result in an accurate model,
  with the right curve

11
Growth or S Curves cont.

• Gompertz

• Logistics (Pearl-Reed)

12
Current work on Qualitative

• Much of current work seems to be addressing
  validity (not repeatable and how to eliminate
  bias)
• Compare and combine with quantitative methods
• Judgmental techniques are used in 40-50 of
  business forecasts
• Tend to be overconfident - but still hits the
  mark in many areas.
• Smaller companies use judgmental techniques more
  often than larger firms