Abstract
A central feature of demand for products in the fashion apparel segment is a pronounced product life cycle—demand for a fashion product tends to rise and fall dramatically in accordance with the rate of public of adoption. Product demands that vary in such a manner can be difficult to forecast, especially in the critical early period of a product’s life, when observed demand can be a very unreliable yardstick of demand later on. This paper examines the applicability of a Bayesian forecasting model—based on one developed for use in the computer industry—to fashion products. To do so, we use an agent-based simulation to produce a collection of demand series consistent with commonly-accepted characteristics of fashion adoption. Using Markov chain Monte Carlo techniques to make predictions using the Bayesian model, we are able quantitatively to demonstrate its superior performance in this application.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
A full summary of the notation in the paper, and a description of the probability distributions used is provided in the appendices.
- 2.
For concreteness’s sake, we will frequently regard one period as 1 day, though we use the term “period” throughout to emphasize the generality of the framework.
- 3.
For technical convenience, we take Purch 0 to be the empty set.
- 4.
Throughout this section, events are assumed (conditionally) independent unless the contrary is noted.
- 5.
i.e. \(i \in \mathtt{Market}\) and \(i\notin \mathtt{Purch}_{t-1}\).
- 6.
Since some of the T j may be zero, we adopt the convention that for any expression •, \(\prod _{j=1}^{0}\mathrm{\bullet } = 1\).:
- 7.
For a comprehensive survey of analytical models of product life cycles, see Mahajan et al. [42].
- 8.
We use the difference between the 20th and 95th percentiles rather than the 95th percentile itself because α j and δ j might reasonably be considered a priori independent, making for easier specification of the model prior.
- 9.
Strictly speaking, truncation on the left should be at a point slightly greater than 0, but the technical elision is of no practical consequence.
- 10.
In this respect, we depart from the precept set out in the introduction requiring us to disassociate the simulation and forecasting model. However, since we do not regard seasonal effects as a defining characteristic of demand for fashion goods (to which life cycle effects are most pivotal), and since alternative approaches to seasonal modeling would be unnecessarily cumbersome here, we consider such a lapse justified.
- 11.
“Boxes” delimit the interquartile range of the distributions, and “whiskers” extend 1.5 times the interquartile range from the ends of the boxes—see Tukey [62] for further details.
- 12.
This is a modification of the more common mean absolute percentage error (MAPE) metric; since the values in the forecast series vary significantly over lifecycle of the corresponding product, the MAPE is apt to distort performance assessment by unduly inflating small absolute errors at either end of the lifecycle.
- 13.
Note that the comparatively poor performance of the ets function here should not be construed as general criticism, but rather as affirmation of the contention that the proposed model is far better suited to the type of series encountered in this context than is a general-purpose forecasting tool.
- 14.
In this respect, the situation here is the obverse of that in [67], where efficacy for a particular manufacturer’s demand series was demonstrated, but more general applicability remained something of an open question.
- 15.
Hines and Bruce [32] point out that “fast fashion” retailers such as Zara and Primark publish revised forecasts on a weekly basis.
References
Agrawal N, Smith SA (1996) Estimating negative binomial demand for retail inventory management with unobservable lost sales. Nav Res Logist 43(6):839–861
Ansari A, Jedidi K, Jagpal S (2000) A hierarchical Bayesian methodology for treating heterogeneity in structural equation models. Mark Sci 19(4):328–347
Armstrong JM (2001) Standards and practices for forecasting. In: Armstrong JM (ed) Principles of forecasting: a handbook for researchers and practitioners. Kluwer, Norwell, pp 171–192
Azoury KS (1985) Bayesian solution to dynamic inventory models under unknown demand distribution. Manag Sci 31:1150–1160
Azoury KS, Miller BS (1984) A comparison of the optimal levels of Bayesian and non-Bayesian inventory models. Manag Sci 30:993–1003
Bass F (1969) A new product growth model for consumer durables. Manag Sci 15:215–227
Bernardo JM, Smith AFM (1994) Bayesian theory. Wiley, Hoboken
Brockwell PJ, Davis RA (2002) Introduction to time series and forecasting, 2nd edn. Springer, New York
Bruce M, Daly L (2007) Challenges of fashion buying and merchandising. In: Hines T, Bruce M (eds) Fashion marketing: contemporary issues, 2nd edn. Elsevier, Oxford, pp 54–69. Chapter 3
Caro F, Gallien J (2007) Dynamic assortment with demand learning for seasonal consumer goods. Manag Sci 53(2):276–292
Choi TM, Li D, Yan HM (2003) Optimal two-stage ordering policy with Bayesian information updating. J Oper Res Soc 54(8):846–859
Choi TM, Li D, Yan HM (2006) Quick response policy with Bayesian information updates. Eur J Oper Res 170(3):788–808
Cohen MD, March JG, Olsen JP (1972) A garbage can model of organizational choice. Adm Sci Q 17:1–25
Cook SR, Gelman A, Rubin DB (2006) Validation of software for Bayesian models using posteior quantiles. J Comput Graph Stat 15(3):675–692
Delre SA, Jager W, Bijmolt THA, Janssen MA (2007) Targeting and timing promotional activities: an agent-based model for the takeoff of new products. J Bus Res 60(8):826–835
Dong X, Manchanda P, Chintagunta PK (2009) Quantifying the benefits of individual-level targeting in the presence of firm strategic behavior. J Mark Res 9(3):301–337
Donohue KL (2000) Efficient supply contracts for fashion goods with forecast updating and two production modes. Manag Sci 46(11):1397–1411
Dvoretzky A, Kiefer J, Wolfowitz J (1952) The inventory problem: case of known distribution of demand. Econometrica 20:187–222
Eppen GD, Iyer AV (1997) Improved fashion buying with Bayesian updating. Oper Res 45(6):805–819
Fox MS, Barbuceanu M, Teigen R (2000) Agent-oriented supply-chain management. Int J Flex Manuf Syst 12:165–188
Gardner ES, McKenzie E (1989) Seasonal exponential smoothing with damped trends. Manag Sci 35(3):372–376
Gelman A (2006) Prior distributions for variance parameters in hierarchical models. Bayesian Anal 1(3):515–533
Gelman A, Hill J (2006) Data analysis using regression and multilevel/hierarchical models. Cambridge University Press, New York
Gelman A, Carlin J, Stern H, Rubin D (2003) Bayesian data analysis, 2nd edn. Chapman & Hall/CRC, Boca Raton
Geweke J (1992) Evaluating the accuracy of sampling-based approaches to calculating posterior moments. In: Bernardo JM, Berger JO, Dawid AP, Smith AFM (eds) Bayesian statistics, vol 4. Clarendon, Wotton-under-Edge/Gloucestershire, pp 169–193
Gilks WR, Best NG, Tan KKC (1995) Adaptive rejection Metropolis sampling within Gibbs sampling. Appl Stat 44:455–472. (Corr: Neal, R.M. 1997 vol 46 pp 541–542)
Gilks WR, Richardson S, Spiegelhalter DJ (1996) Markov chain Monte Carlo in practice. Chapman & Hall, Boca Raton
Goldenberg J, Han S, Lehmann DR, Hong JW (2009) The role of hubs in the adoption process. J Mark 73(2):1–13
Griffiths W (2004) A Gibbs sampler for the parameters of a truncated multivariate normal distribution. In: Becker R, Hum S (eds) Contemporary issues in economics and econometrics: theory and application, Edward Elgar, Cheltenham/Northampton, pp 75–91
Gurnani H, Tang C (1999) Note: optimal ordering decisions with uncertain cost and demand forecast updating. Manag Sci 45(10):1456–1462
Harvey AC (1989) Forecasting, structural time series models and the Kalman filter. Cambridge University Press, Cambridge
Hines T, Bruce M (2007) Fashion marketing. Elsevier, Oxford
Hunter NA, Valentino P (1995) Quick response—ten years later. Int J Cloth Sci Technol 7:30–40
Hyndman RJ, Khandakar Y (2008) Automatic time series forecasting: the forecast package for R. J Stat Softw 27(3):1–22. http://www.jstatsoft.org/v27/i03
Hyndman RJ, Koehler AB (2006) Another look at measures of forecast accuracy. Int J Forecast 22:679–688
Hyndman RJ, Koehler AB, Snyder RD, Grose S (2002) A state space framework for automatic forecasting using exponential smoothing methods. Int J Forecast 18(3):439–454
Iglehart D (1964) The dynamic inventory problem with unknown demand distribution. Manag Sci 10:429–440
Iyer A, Bergen ME (1997) Quick response in manufacturer-retailer channels. Manag Sci 43(4):559–570
Kang KY (1999) Development of an assortment-planning model for fashion-sensitive products. PhD thesis, Virginia Polytechnic Institute and State University
Kiesling E, Gunther M, Stummer C, Wakolbinger LM (2009) An agent-based simulation model for the market diffusion of a second generation biofuel. In: Winter simulation conference, Austin
Krishnamoorthy K (2006) Handbook of statistical distributions with applications. Statistics: a series of textbooks and monographs. Chapman and Hall/CRC, Boca Raton
Mahajan V, Muller E, Wind Y (eds) (2000) New-product diffusion models. Kluwer, Norwell
McCool JI (2012) Using the Weibull distribution: reliability, modeling, and inference. Wiley, Hoboken
Miller B (1986) Scarf’s state reduction method, flexibility and a dependent demand inventory model. Oper Res 34(1):83–90
Moe WW, Fader PS (2002) Using advance purchase orders to forecast new product sales. Mark Sci 21(3):347–364
Müller P (1991) A generic approach to posterior integration and Gibbs sampling. Technical report 91-09, Department of Statistics, Purdue University
Murray GR Jr, Silver EA (1966) A Bayesian analysis of the style goods inventory problem. Manag Sci 12(11):785–797
Narayanan S, Manchanda P (2009) Heterogeneous learning and the targeting of marketing communication for new products. Mark Sci 28(3):424–441
North MJ, Macal CM, St Aubin J, Thimmapuram P, Bragen M, Hahn J (2010) Multi-scale agent-based consumer market modeling. Complexity 15(5):37–47
Pan R, Peng Y, Ding Z (2006) Belief update in Bayesian networks using uncertain evidence. In: International conference on tools with artificial intelligence, ICTAI, Washington, DC, pp 441–444
Pearl J (1988) Probabilistic reasoning in intelligent systems: networks of plausible inference. Morgan Kaufmann, San Mateo
Petris G (2010) dlm: Bayesian and likelihood analysis of dynamic linear models. http://cran.r-project.org/web/packages/dlm/index.html
Rand W, Rust RT (2011) Agent-based modeling in marketing: guidelines for rigor. Int J Res Mark 28(3):181–183
Robert CP, Casella G (2005) Monte Carlo statistical methods, 2nd edn. Springer, New York
Rossi PE, McCulloch RE, Allenby GM (1996) The value of purchase history data in target marketing. Mark Sci 15(4):321–340
Rossi PE, Allenby GM, McCulloch RE (2005) Bayesian statistics and marketing. Wiley, Hoboken
Scarf H (1959) Bayes solutions of the statistical inventory problem. Ann Math Stat 30(2):490–508
Schenk TA, Loffler G, Rauh J (2007) Agent-based simulation of consumer behavior in grocery shopping on a regional level. J Bus Res 60(8):894–903
Schwoon M (2006) Simulating the adoption of fuel cell vehicles. J Evol Econ 16(4):435–472
Sharif MN, Islam MN (1980) The Weibull distribution as a general model for forecasting technological change. Technol Forecast Soc Change 18:247–256
Sproles GB (1981) Analyzing fashion life cycles: principles and perspectives. J Mark 45:116–124
Tukey J (1977) Exploratory data analysis. Addison-Wesley, Reading
Valtorta M, Young-Gyun K, Vomlel J (2002) Soft evidential update for probabilistic multiagent systems. Int J Approx Reason 29:71–106
Walsh W, Wellman M (1999) Modeling supply chains formation in multi-agent systems. In: IJCAI-99 workshop agent-mediated electronic commerce, Stockholm
Watts DJ, Strogatz SH (1998) Collective dynamics of ‘small-world’ networks. Nature 393:409–410
Wu DS, Kempf KG, Atan MO, Aytac B, Shirodkar SA, Mishra A (2010) Improving new product forecasting at Intel corporation. Interfaces 40(5):385–396
Yelland PM (2010) Bayesian forecasting of parts demand. Int J Forecast 26(2):374–396
Yelland PM, Kim S, Stratulate R (2010) A Bayesian model for sales forecasting at Sun Microsystems. Interfaces 40:118–129
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Yelland, P.M., Dong, X. (2014). Forecasting Demand for Fashion Goods:A Hierarchical Bayesian Approach. In: Choi, TM., Hui, CL., Yu, Y. (eds) Intelligent Fashion Forecasting Systems: Models and Applications. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-39869-8_5
Download citation
DOI: https://doi.org/10.1007/978-3-642-39869-8_5
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-39868-1
Online ISBN: 978-3-642-39869-8
eBook Packages: Business and EconomicsBusiness and Management (R0)