Accurate forecasts of sales are crucial to most companies. To produce their forecasts many companies use a combination of statistical methods and management judgment. Ideally, these methods should complement each other, but research has shown that management judgment is often subject to biases, is used inconsistently and is often applied when it is not appropriate.
My research is aimed at designing forecasting software that will provide support to managers in their use of judgment so that these problems are reduced and greater forecast accuracy can be achieved. Why do managers frequently adjust or replace the forecasts of reliable statistical methods? How can managers be encouraged to refrain from making judgmental interventions when these are likely to reduce accuracy? When the use of judgment is appropriate, how can managers be supported in making their judgments? The second theme of my research is a broad theme which includes providing support to managers facing decisions involving multiple objectives and integrating scenario planning with multiattribute decision methods.