When it comes to forecasting or decision-making, relying on the arithmetic mean, or average, can lead you to a seriously flawed plan of attack, according to the Management Insights feature in the current issue of Management Science, the flagship journal of the Institute for Operations Research and the Management Sciences (INFORMS®).
Management Insights, a regular feature of the journal, is a digest of important research in business, management, operations research, and management science. It appears in every issue of the monthly journal.
"Why Intuitive Averages Can Make Bad Forecasts and Counter-Intuitive Forecasts Can Do Better" is by Steven Shugan and Debanjan Mitra of the University of Florida.
Although some would argue, for example, that the best predictor of sales for a vocal-artist's next album is the average of the artist's previous album sales, the authors argue that this type of reliance on averages is wrong. Although it is natural to use the artist's average sales as a metric to predict future sales, it turns out that the artist's maximum sales may be the better predictor, even if some of the artist's previous albums were terrible.
Similarly, it may be more desirable to invest with an entrepreneur who has had one highly successful startup and several failures rather than to invest with an entrepreneur who on average had more success.
The authors explain that sometimes rare events – for example, a success in an environment where failure is common – sometimes convey more information and averaging tends to ignore that critical information. The authors show that this counter-intuitive result can apply in many situations, including, among others, the performance of professional sports players, the box office draw of actors, and the research productivity of scientists. In addition to explaining why nonaveraging metrics (such as the maximum) can perform better, they provide important guidelines for decision makers.
Source: Institute for Operations Research and the Management Sciences