Abstract: Systems and methods for forecasting demand for a test book. One method according to the present invention includes dividing books into a plurality of categories with each of the categories having a unifying concept, developing a demand forecast model for each category by selecting a plurality of variables and, for each category, assigning each selected variable a value based upon historical data showing the importance of each selected variable in predicting demand for books in the category. The test book is assigned to one of the plurality of categories based upon a similarity between the test book and the unifying concept of the assigned category. A target audience is provided with the test book and the audience is surveyed to capture data that can be used to calculate a forecasted value for each selected variable in the assigned category. The forecasted value is then used for each selected variable in the demand forecast model for the assigned category to forecast a demand for the test book.