On the differential benchmarking of promotional efficiency with machine learning modeling (I): Principles and statistical comparison
Highlights
► Empirical models of sales promotion are relevant for marketing strategies. ► A simple statistical tool allows operative comparisons among promotional models. ► Bootstrap statistical description is used to evaluate the models in terms of average and scatter measurements. ► Different figure of merits, and structured parameter selection, allowed an optimized promotion modeling. ► Prediction quality was robust with respect to the design parameters selection.
Introduction
The current economic landscape, characterized by financial instability and the consequent changes in consumer behavior, is driving a transformation in food retailer decision, bringing to a new and more aggressive promotional perspective (Quelch, 2008). As an example of this situation, the dramatic sales reduction of food products in Spain, which has led retailers in the industry to implement new approaches, such as the intense use of private label products, can be mentioned. In addition, it has been also searched to increase consumer’s frequent purchases through promotional activities, such as promotional discounts, feature advertising, and promotional packs (e.g., “buy 3 get 1 free”) (Quelch, 2008). Therefore, there is no denying that sales promotion has become in recent years a key tool for marketing strategies in retail food markets, and for this reason, investment in this area has strongly increased, reaching values over 50% of marketing budgets in relation to other communications tools (Villalba & Iñaki, 2002).
The present economic situation, along with food retailer’s strategies and increasing investment in promotional activities, has motivated an important number of research efforts to characterize sales promotion and to measure promotional efficiency. Existing models for analysing sales promotions effects can be classified into two separate groups. In the first group, namely theoretical models, consumer behavior is basically evaluated considering a sociological and psychological perspective, whilst in the second group of empirical models, promotional structures based on empirical information extracted from historical databases are usually built.
Within that last group, the efforts have been focused during the last decades on the understanding of sales promotion dynamics based on classical statistical analysis methods, and more recent works are concentrated towards the machine learning algorithmic and data mining techniques, as powerful tools to extract information from existing recorded data (Mitchell, 1997, Van Heerde et al., 2000). Machine learning techniques aim to find recurring patterns, trends, or rules, which can explain the data behavior in a given context, and then allows to extract new knowledge on the consumer behavior, to improve the performance of marketing operations, and to estimate the commonly called Deal Effect Curve (DEC). In particular, a vast amount of knowledge has been extracted from machine learning techniques, although not all promotional behaviors have been studied and there is still room for in depth further studies (Bell et al., 1999, Blatterg et al., 1995, Leeflang and Wittingk, 2000). More specifically, operational problems arise in machine learning promotional modeling, when based on nonlinear estimation techniques, for evaluating and demonstrating working hypothesis (Liu et al., 2004, Martínez-Ruiz et al., 2005, Martı´nez-Ruiz et al., 2006b, Martínez-Ruiz et al., 2006a, Van Heerde et al., 2001, Wang et al., 2008). In this paper we present, prior to an in detail analysis of machine-learning performance, a pre-evaluation of different figures of merit, to asses their impact on the final result, together with an in depth analysis for design parameters selection. Results obtained here allowed establishing and validating figures of merit and the design parameters selection procedure to be applied for pricing promotion study. Specifically, we applied these results in the companion paper (Soguero-Ruiz, Gimeno-Blanes, Mora-Jiménez, Martı´nez-Ruiz, & Rojo-Álvarez, 2012), to evaluate four well-known machine learning algorithms in two real databases for two categories presenting dramatically different promotional behavior.
The draw of the paper is as follows. Section 2 includes a review of basic concepts of sales promotion in retailer environments, as well as a summary of previous work on machine learning in the context of marketing. Section 3 gives a short description of nonparametric inference paired hypothesis tests, based on bootstrap resampling, and actual risk evaluation of a set of adequate figures of merit is introduced. The figure of merit benchmark represents a relevant contribution of this paper, as it ends up becoming an operative tool for decision support in promotional modeling with machine learning techniques. Section 4 summarizes experiments and results based on real data. Finally in Section 5, conclusion and remarks are presented.
Section snippets
Background
Though many definitions have been published for the term sales promotion (Blattberg and Neslin, 1990, Kotler and Keller, 2005, Yeshin, 2006), none of them are generally accepted, but general consensus suggests that sales promotions consist basically of short-time sales incentives (Blattberg and Neslin, 1990, Kotler and Keller, 2005). For instance, the American Marketing Association (AMA) defines sales promotion as a media and non media marketing pressure applied for a predetermined, limited
Paired bootstrap resampling of actual risk
In this section, we present the data model for sales promotion, the statistical learning from samples techniques used in this study to design the sales promotion model, and the design parameters selection that has to be addressed usually by means of cross-validation techniques of some kind, which raises the concepts of empirical risk and actual risk. Additionally, different figures of merit are presented, as well as the Bootstrap resampling technique, to yield an estimation of the actual risk.
Experiments
This section presents the first set of experimental studies, using two food products databases, namely, milk and beer categories. In Experiment 1, some examples of the estimation of the own-effect given by the DEC in the absence of other explanatory effects were checked, for evaluating the necessity of considering additional exogenous effects in order to provide with an adequate promotional sales model. Then, the relevant issue of the design parameter selection is analyzed for the proposed
Conclusions
This work first aimed to analyze the multiple and simultaneous effect of different promotional activities from retailers with individual DEC by products. It was subsequently reported the complexity of the cross-effects among different promotional activities taking place at the same time. These promotional activities and discounts strategies represent a second-order effect of the current economic situation where retailers are seeking for strategies that lead to better performance – especially in
Acknowledgment
This work was supported by Fundación Ramón Areces (Spain).
References (41)
- et al.
Chain-wide and store-level analysis for cross-category management
Journal of Retailing
(2007) - et al.
Assessing the competitive impact of type, timing, frequency, and magnitude of retail promotions
Journal of Business Research
(1997) - et al.
Building models for marketing decisions: Past, present and future
International Journal of Research in Marketing
(2000) - et al.
Decomposing the sales promotion bump accounting for cross-category effects
International Journal of Research in Marketing
(2008) - et al.
Using daily store-level data to understand price promotion effects in a semiparametric regression model
Retailing and Consumer Services
(2006) - et al.
Relating price sensitivity to retailer promotional variables and pricing policy: An empirical analysis
Journal of Retailing
(1996) - et al.
Exploring the effect of retail sector and firm characteristics on retail price promotion strategy
Journal of Retailing
(2003) Stochastic parsing and evolutionary algorithms
Applied Artificial Intelligence
(2009)- et al.
The descomposition of promotional response: an empirical generalization
Marketing Science
(1999) - et al.
The effects of reference prices in retail advertisements
The Journal of Marketing
(1981)
Sales promotion: Concepts, methods, and strategies
How promotion work
Marketing Science
New figures of merit for best-first probabilistic chart parsing
Computational Linguistics
Nearest neighbor pattern classification
IEEE Transactions on Information Theory
Pattern classification
An introduction to the bootstrap
Marketing for the entrepreneur: Customer focus to multiple constituencies
Entrepreneurship and Economic Growth in the American Economy
Neural networks
Cited by (14)
Dimensionality reduction and ensemble of LSTMs for antimicrobial resistance prediction
2023, Artificial Intelligence in MedicineInformative variable identifier: Expanding interpretability in feature selection
2020, Pattern RecognitionCitation Excerpt :Special attention has also been lately paid to techniques that assess a feature’s relevance by analyzing the confidence level or consistency of a weight assigned to the feature and estimated with resampling or subsampling techniques. These methods have been proven useful for identification of relevant brain regions or for mental disease biomarkers design [8,15], for block analysis of best student profiles in a worldwide study such as the Programme for International Student Assessment (PISA) [16], for prediction of infection [17], or for forecasting promotional sales [18,19]. The IVI feature selection algorithm proposed here fits into EFS by the ensemble, as far as its main target is to provide us with a better knowledge of the relationships among features.
Statistical nonlinear analysis for reliable promotion decision-making
2014, Digital Signal Processing: A Review JournalCitation Excerpt :Statistical learning for promotional modeling. Other researchers have proposed nonlinear statistical learning algorithms, including such classic nonparametric methods as k nearest neighbors (k-NN) and kernel estimators [34], as well as the new learning techniques such as neural networks and support vector machines [12,13,35–38]. It is worthwhile to note that nonlinearity, nonnormal errors, and heteroscedasticity are automatically harmonized by these kinds of methods.
On the differential benchmarking of promotional efficiency with machine learning modelling (II): Practical applications
2012, Expert Systems with ApplicationsCitation Excerpt :The draw of the paper is as follows. Section 2 presents a description of the machine learning techniques analyzed in this work, and a short summary of the method proposed in the companion paper (Soguero-Ruiz et al., 2012). Afterward, the two databases to be used for sales promotion modelling are described (milk and beer category) in Section 3.
Marketing Strategy and Artificial Intelligence State of the Art and Research Agenda
2024, Journal of Telecommunications and the Digital EconomyA Subtle Design of Prediction Models Using Machine Learning Algorithms for Advocating Selection and Forecasting Sales of Garments: A Case Study
2024, Lecture Notes in Networks and Systems