Optimism bias and differential information use in supply chain forecasting

Companies tend to produce over optimistic forecasts of demand. Psychologists have suggested that optimism, at least in part, results from selective bias in processing information. When high demand is more desirable than low demand information favouring high demand may carry more weight than information that suggests the opposite. To test this, participants in an experiment used a prototypical forecasting support system to forecast the demand uplift resulting from sales promotion campaigns. One group was rewarded if demand exceeded an uplift of 80%. All of the participants were also supplied with information some of which supported an uplift of greater than 80% and some of which suggested that the uplift would fall below this value. This enabled an assessment to be made of the extent to which those who were rewarded if the uplift exceeded 80% paid more attention to the positive information than those who were not rewarded. While the experiment provided strong evidence of desirability bias there was no evidence that the group receiving rewards for sales exceeding the threshold made greater use of positive reasons. Possible explanations for this are discussed, together with the implications for forecasting support system design.