Effective judgmental forecasting in the context of fashion products⋆

• Published date: 01 May 2015
• Date: 01 May 2015
• Author: Andreas B. Eisingerich,Enno Siemsen,Allègre L. Hadida,Matthias Seifert
• DOI: http://doi.org/10.1016/j.jom.2015.02.001

Journal of Operations Management 36 (2015) 33–45

Contents lists available at ScienceDirect

Journal of Operations Management

journal homepage: www.elsevier.com/locate/jom

Effective judgmental forecasting in the context of fashion products夽

Matthias Seiferta,∗, Enno Siemsen b, Allègre L. Hadidac, Andreas B. Eisingerich d

aIE Business School, IE University, Spain

bCarlson School of Management, University of Minnesota, United States

cUniversity of Cambridge Judge Business School & Magdalene College, United Kingdom

dImperial College Business School, Imperial College London, United Kingdom

article info

Article history:

Received 4 June 2014

Received in revised form 7 February 2015

Accepted 12 February 2015

Available online 28 February 2015

Accepted by Thomas Younghoon Choi

Keywords:

Judgmental forecasting

Fashion products

Lens model design

Demand uncertainty

Music industry

New product forecasting

abstract

We study the conditions that influence judgmental forecasting effectiveness when predicting demand in

the context of fashion products. Human judgment is of practical importance in this setting. Our goal is

to investigate what type of decision support, in particular historical and/or contextual predictors, should

be provided to human forecasters to improve their ability to detect and exploit linear and nonlinear

cue–criterion relationships in the task environment. Using a field experiment on new product forecasts

in the music industry, our analysis reveals that when forecasters are concerned with predictive accuracy

and only managerial judgments are employed, providing both types of decision support data is benefi-

cial. However, if judgmental forecasts are combined with a statistical forecast, restricting the decision

support provided to human judges to contextual anchors is beneficial. We identify two novel interac-

tions demonstrating that the exploitation of nonlinearities is easiest for human judgment if contextual

data are present but historical data are absent. Thus, if the role of human judgment is to detect these

nonlinearities (and the linearities are taken care of by some statistical model with which judgments are

combined), then a restriction of the decision support provided makes sense. Implications for the theory

and practice of building decision support models are discussed.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

The accurate prediction of the commercial success of newly

launched products or services represents a crucial managerial

problem (Steenkamp et al., 1999; Stremersch and Tellis, 2004;

Van den Bulte and Stremersch, 2004). Generating such forecasts

can be extremely difficult, particularly in environments involv-

ing fashion-oriented consumer products (hereafter referred to as

“fashion products”), where the nature of the products may con-

tain a substantial creative, artistic component, and consumer taste

constantly changes (Christopher et al., 2004; Hines and Bruce,

2007; Hirsch, 1972). In more conventional forecasting domains, for

example when predicting the demand of machine spare parts in

manufacturing (e.g., Sani and Kingsman, 1997) or when estimating

electricity demand (e.g., Taylor and Buizza, 2003), large amounts

of historical data are often available to calibrate decision support

夽Please note that an article based on portions of the same empirical data sample

has appeared in Organizational Behavior and Human Decision Processes (Volume

120, January 2013). Cross-references are used where appropriate.

∗Corresponding author. Tel.: +34 915689600.

E-mail address: matthias.seifert@ie.edu (M. Seifert).

models and achieve high levels of model accuracy. Forecasts about

the demand of fashion products, on the other hand, often lack

such integral information as the demand pattern tends to be highly

uncertain (Choi et al., 2014; Green and Harrison, 1973; Sichel, 2008;

Sun et al., 2008).

Companies producing fashion products for which consumer

tastes and preferences cannot be tracked continuously often

perceive themselves more as trend-setters than trend-followers

(Eliashberg et al., 2008). For instance, in the apparel industry, firms

face the challenge of quickly commercializing new designs that are

introduced during the New York or Paris Fashion Week in order

to create and satisfy new consumer demand. Yet, forecasts about

futuresales of new clothing designs are highly volatile. They depend

both on managers’ ability to accurately anticipate uncertain con-

sumer preferences and on their firm’s time-to-market capability

relative to its competitors. In such supply-driven environments

(Moreau and Peltier, 2004), conventional time series methods typ-

ically cannot be employed to predict demand with reasonable

accuracy (Eliashberg and Sawhney, 1994; Moe and Fader, 2001;

Sawhney and Eliashberg, 1996). Instead, researchers have proposed

several approaches to overcome the problem of model calibra-

tion when only limited and/or unreliable data are available (for

an extensive review, please see Nenni et al., 2013). In particular,

http://dx.doi.org/10.1016/j.jom.2015.02.001

0272-6963/© 2015 Elsevier B.V. All rights reserved.