Determining the appropriate amount of data for classifying consumers for direct marketing purposes

Heilman, Carrie M.; Kaefer, Frederick; Ramenofsky, Samuel D.

doi:10.1002/dir.10057

Cited by 26 publications

(10 citation statements)

References 59 publications

(79 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We contribute to this literature by adopting a multiple-campaign orientation that considers both cross-sectional and temporal heterogeneity in response. Our work is also related to a body of work focused on using customer information to target customers (Heilman et al 2003, Rossi et al 1996. Our work adds to this literature by quantifying the value of tracking and acting on both unobserved preference heterogeneity and observed transaction history measures.…”

Section: Introductionmentioning

confidence: 95%

Dynamic Customer Management and the Value of One-to-One Marketing

2009

View full text Add to dashboard Cite

The concept of one-to-one marketing is intuitively appealing, but there is little research that investigates the value of individual-level marketing relative to segment-level or mass marketing. In this paper, we investigate the financial benefits of and computational challenges involved in one-to-one marketing. The analysis uses data from an online grocery and drug retailer. Like many retailers, this firm uses multiple promotional instruments including discount coupons, free shipping offers, and a loyalty program. We investigate the impact of customizing these promotions on the two most important consumer decisions: the decision to buy from the store and expenditure. Our modeling approach accounts for two sources of heterogeneity in consumers' responsiveness to various marketing mix elements: cross-sectional differences across consumers and temporal differences within consumers based on the purchase cycle. The model parameter estimates are fed into a dynamic programming model that determines the optimal number, sequence, and timing of promotions to maximize retailer profits. A series of policy simulations show that customizing promotions leads to a significant increase in profits relative to the firm's current practice of uniform promotions. However, the effectiveness of various promotions varies because of both cross-sectional differences in consumers as well within consumer heterogeneity due to purchase cycle factors. For instance, we find that free shipping tends to be the preferred instrument for re-acquiring lapsed customers, whereas an across-the-board price cut (via a discount coupon) is the most effective tool for managing the segment of most active customers. Interestingly, we find that customizing based on within-customer temporal heterogeneity contributes more to profitability than exploiting variations across consumers. This is important because the computational burden of implementing the dynamic optimization to account for cross-sectional heterogeneity is far greater than accounting for temporal heterogeneity. Furthermore, targeting promotions based only on timing rather than the nature and magnitude of the offers across consumers alleviates the public relations risks of price discrimination. Implications for marketing managers are also discussed.CRM, one-to-one marketing, promotions, dynamic programming, heterogeneity, database marketing

show abstract

Section: Introductionmentioning

confidence: 95%

Dynamic Customer Management and the Value of One-to-One Marketing

2009

View full text Add to dashboard Cite

show abstract

“…To illustrate, Rossi, McCulloch, and Allenby (1996) used short purchase histories to develop a model that increased the redemption rate of coupons by 2.5 times the rate for blanket couponing. Heilman, Kaefer, and Ramenofsky (2003) build on the work of Rossi et al by examining the effects of having more or less past-purchase data. Similar models are offered for predicting preferences for Japanese-made cars (Yang & Allenby, 2003) and identifying households that spent large amounts on grocery shopping (Bodapati & Gupta, 2004).…”

Section: Behavioral Scoring Modelsmentioning

confidence: 99%

Customer Behavior in an Online Ordering Application: A Decision Scoring Model*

Boyer¹,

Hult²

2005

Decision Sciences

102

View full text Add to dashboard Cite

This research presents the development of behavioral scoring models to predict future customer purchases in an online ordering application. Internet retailing lowers many barriers for customers switching between retailers for repeat purchases; thus, retaining existing customers is a key challenge for achieving profitability. Survey data were collected from 1,089 online customers of two companies. The subjective survey data were then used to predict purchases over the ensuing 12 months based on data from the company databases. The analysis illustrates the general applicability of predictive models of future customer purchases while also demonstrating the need to develop specific models tailored for an individual company's operating and marketing environment. The models provide insight on how companies can target marketing dollars more effectively and allocate investment across multiple operational areas for maximum return. The research answers a call for rigorous research in the area of predictive marketing, an area in which many companies are excelling but where there is a scarcity of detailed knowledge regarding application of such models.

show abstract

“…Classification with LDA involves classifying subjects into one of several groups on the basis of a set of measurements. LDA has a long tradition in the marketing literature for providing solutions to problems involving discrete outcomes, such as choice or classification (Heilman et al, 2003) and for its competency in predicting choice as a function of past behavior (Mela et al, 1997). LDA assumes certain statistical characteristics of the data, such as multivariate normality and homogeneity of variance/covariance matrices.…”

Section: Discriminant Analysismentioning

confidence: 99%

Profiling organ donors in Egypt using intelligent modeling techniques

Mostafa

2008

Marketing Intelligence &Amp; Planning

View full text Add to dashboard Cite

Purpose -This study uses intelligent modeling techniques with the purpose of examining the effect of various demographic, cognitive and psychographic factors on organ donation in Egypt. Design/methodology/approach -Two artificial neural network models (multi-layer perceptron neural network and probabilistic neural network) are compared to two standard statistical methods (linear discriminant analysis and logistic regression). The variable sets considered are sex, age, educational level, religion, altruistic values, perceived benefits/risks of organ donation, organ donation knowledge, attitudes toward organ donation, and intention to donate organs. Findings -The results show that artificial neural networks outperform traditional statistical techniques in profiling potential organ donors due to their robustness and flexibility of modeling algorithms. Originality/value -The paper shows how it is possible to identify various dimensions of organ donation behavior by uncovering patterns in the dataset, and also shows the classification abilities of two neural network techniques.

show abstract

Determining the appropriate amount of data for classifying consumers for direct marketing purposes

Cited by 26 publications

References 59 publications

Dynamic Customer Management and the Value of One-to-One Marketing

Dynamic Customer Management and the Value of One-to-One Marketing

Customer Behavior in an Online Ordering Application: A Decision Scoring Model*

Profiling organ donors in Egypt using intelligent modeling techniques

Contact Info

Product

Resources

About