Fader_hardie_interf_01 - Forecasting Repeat Sales at CDNOW:...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
Copyright q 2001 INFORMS 0092-2102/01/3103/S094/$05.00 1526–551X electronic ISSN MARKETING—BUYER BEHAVIOR FORECASTING—APPLICATIONS This paper was refereed. INTERFACES 31: 3, Part 2 of 2, May–June 2001 (pp. S94–S107) Forecasting Repeat Sales at CDNOW: A Case Study Peter S. Fader [email protected] The Wharton School University of Pennsylvania Philadelphia, Pennsylvania 19104-6371 Bruce G. S. Hardie [email protected] London Business School Regent’s Park London NW1 4SA, United Kingdom We conducted a modeling exercise in conjunction with the on- line music retailer CDNOW to develop a simple stochastic model of buyer behavior capable of forecasting medium-term aggregate CD purchasing by a cohort of new customers. We modeled weekly sales using a ±nite mixture of beta-geometric distributions with a separate time-varying component to cap- ture nonstationarity in repeat buying. The resulting model can easily be implemented within a standard spreadsheet environ- ment (for example, Microsoft Excel). It does a good job of de- scribing the underlying sales patterns and produces an excel- lent medium-term forecast. W ith the growth of e-commerce, many companies are facing chal- lenges in ±guring out how to make effec- tive and ef±cient use of the detailed trans- action information that they are rapidly accumulating. While some writers on data- base marketing and one-to-one marketing suggest using statistical models to gain managerial insights [Mulhern 1999; Forres- ter Report 1999], there are very few pub- lished examples to guide managers’ efforts in this direction. We undertook an exploratory study in conjunction with CDNOW, a leading on- line music retailer, to develop an easily implementable model of buyer behavior capable of forecasting medium-term ag- gregate CD purchasing by a cohort of CDNOW customers. Aggregate-level fore- casts are critical inputs to any attempt to value a customer base, and they serve as a diagnostic to help ±rms gauge the effec-
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
CDNOW May–June 2001 S95 tiveness of various short-term marketing programs (for example, they can provide baseline sales estimates against which the performance of promotions can be evaluated). At the time of this study (1997 through 1998), many commentators felt that the Internet was still in its infancy and that any forecasting exercise would be futile. For example, Buchanan and Lukaszewski [1997, p. 143] made the comment that: At this stage of the Internet’s evolution, accu- rate sales forecasts are as much of an oxy- moron as “military intelligence.” However, it was—and still is—our view that patterns of buying behavior are fairly consistent across purchasing channels (in- cluding the Internet) and that developing a forecasting model would therefore be a fruitful exercise. Background CDNOW is one of the oldest and largest online retailers, having sold different forms of music (and related products) on the World Wide Web since 1994. It carries approximately 500,000 different albums— about 10 times as many as the typical bricks-and-mortar megastore—and it re- ports store traf±c of over 200,000 visitors per day. During its ±rst ±ve years of
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 12/27/2011 for the course BA 111 taught by Professor Hothanhphong during the Spring '11 term at Chaminade University.

Page1 / 14

Fader_hardie_interf_01 - Forecasting Repeat Sales at CDNOW:...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online