ORIGINAL PAPER An analytic network process approach for sales forecasting Dimitra Voulgaridou Æ Konstantinos Kirytopoulos Æ Vrassidas Leopoulos Received: 24 October 2008 / Accepted: 24 October 2008 / Published online: 13 November 2008 Ó Springer-Verlag 2008 Abstract New product sales forecast is a difficult but essential task for companies, as it drives a variety of multifunctional decisions. Most new product forecasting literature presents sophisticated statistical techniques and complex methodologies with little or no thought given to ease of implementation. As a result, there is a significant gap between what forecast academics create and what forecast models professionals really seek. This study proposes a new product sales forecast model based on an Analytic Network Process framework. The analytic network process (ANP) is a multicriteria decision analysis technique that takes into account both tangible and intangible criteria with interdependencies and feedback among them. It is argued that the proposed framework is more flexible and is more comprehensive than traditional methods and previously developed models. The ANP model is illustrated through a real example concerning the sales forecast of a new edition (book). Keywords Analytic network process  New product sales forecast  Publishing industry  Multicriteria decision analysis D. Voulgaridou (&)  V. Leopoulos Mechanical Engineer Department, National Technical University of Athens, Iroon Polytechniou 9, 15780 Athens, Greece e-mail: dvoulg@central.ntua.gr V. Leopoulos e-mail: vleo@central.ntua.gr K. Kirytopoulos Financial and Management Engineering Department, University of the Aegean, Fostini 31, 82100 Chios, Greece e-mail: kkir@central.ntua.gr 123 Oper Res Int J (2009) 9:35–53 DOI 10.1007/s12351-008-0026-2