Designing an intelligent decision support system for effective negotiation pricing: A systematic and learning approach
Introduction
Intelligent negotiation pricing and differential pricing are prevalent in retailing and business-to-business (B2B), and it is playing an increasingly important role in electronic businesses. In traditional retailing, it is natural to provide standard products and services to all customers at a standard price. In recent years, with the rapid development of marketing science, it is well recognized that different marketing and pricing strategies should be applied to different segments of customers, because differential pricing can substantially enhance organizational profitability and improve customer satisfaction. This emerging phenomenon often occurs in the purchase of expensive products/services (e.g., cars, houses, and systems). In particular, tailor-made products/services require a sales process to negotiate and settle the final price with customers individually. Furthermore, the outcomes of negotiation pricing and differential pricing often have a long-term impact on the organizational supply chain relationship and the reputation of business in the B2B arena. In supply chain management, for instance, price negotiation takes place during annual price reviews, thereby providing an opportunity for suppliers to adjust prices in response to recent changes in costs, as well as the customer relationships [20], [21]. Many examples of this scenario can be found in the industry. For example, broadband wireless service providers may provide differentiated services to their customers with a variety of customized prices [9]; theaters provide personalized prices of last-minute tickets depending on the time remaining and the customer's location [15].
To increase the effectiveness and efficiency of negotiation pricing, companies often delegate certain degrees of pricing authority to sales representatives who have direct contact with and better knowledge of customers. The agreed price will then be approved by the pricing manager. However, the salespeople may have different sales skills and preferences, as empirical findings have revealed that sales representatives might offer too many price concessions in order to ensure the order [30]. Therefore, from the perspective of pricing managers, disclosing the reservation price to all human agents is unfavorable. In essence, decision-making for negotiation pricing is a rather complicated process because it needs to consider information from a plethora of organizational dimensions to identify the right price for the company. It is worth noting that negotiation pricing does not support the entire dynamic and interactive process of price negotiation. Instead, it supports the most fundamental problem in the price negotiation process, which helps pricing managers to identify the right price when offering the unique product/service to each individual customer.
Prior research efforts have been devoted to providing decision support for negotiation pricing through different techniques, including game theory (GT) [9], [34], neural networks [2], [20], expert systems [4], case-based reasoning [15], and fuzzy logic [12], [17], [18]. Notwithstanding, certain drawbacks and challenges still exist with these approaches because some hypothetical assumptions are difficult to achieve in real scenarios, and the assessment of utility functions is not feasible in many studies due to heterogeneity and incidental parameters problems. Expert systems are heavily reliant on expert knowledge and/or static negotiation strategies. Hence, capturing knowledge in manual ways for the resultant system would be less flexible and inefficient to handle the dynamic changes and new cases in negotiations. Similarly, the neural network approach [2], [20] is limited in its interpretability, and its derived results are questionable for the end users.
These extant approaches should resort to adequate and precise information provided by negotiation parties for decision making. However, uncertain information is often inherent in the dynamic negotiation environments of real negotiation scenarios. The involvement of uncertain information is an extremely imperative but often under-addressed issue in negotiation pricing. Fuzzy set theory [32] is well regarded as a useful tool for handling uncertain information, and preserving transparency and interpretability in modeling. Recently, several research efforts [12], [17], [18], [34] have employed this technique to provide decision support for negotiation pricing. In essence, the majority of the existing approaches either focus on the representation of involved uncertain attributes by using linguistic terms, or employ expert knowledge to build static fuzzy rule-based systems for reasoning. Yet, slight changes in negotiation conditions may necessitate substantial expert interventions to modify the corresponding rules to reflect the new conditions. Moreover, it is difficult to validate and assess the quality of knowledge captured from experts. Therefore, it is more desirable to build a fuzzy system (FS) that would automatically learn from historical records to generate the fuzzy rules, rather than completely depending on external knowledge. Additionally, when the number of influential factors is large, a standard FS easily suffers from the problem of dimensionality, since the number of required modeling parameters and fuzzy rules exponentially increases with the number of involved attributes. As such, dealing with uncertain information within negotiation pricing is of particular interest to researchers and practitioners. Given the different features (e.g., dimensionality and data coverage) of the available historical dataset, choosing the most appropriate FSs is another crucial task faced by the end users of decision support systems (DSSs).
In an effort to remedy these pressing issues and challenges, this study substantially extends the initial work of [8], and proposes a systematic and learning approach to provide decision support for negotiation pricing through FS theory. In essence, this study concerns bilateral negations on the price, and considers the negotiation pricing problem particularly from the seller's point of view to provide intelligent decision support for pricing managers. Given a set of historical records, mathematical relationships between influential factors and the proposed price will be built by both learning from the data itself and integrating expert knowledge. It is believed that the proposed model can be leveraged to better predict the will-to-pay and other reference prices (e.g., reservation price, target price, and initial price) for unforeseeable transactions. Beyond the simplified FS with a single input and a single output module (SFS-SISOM) that has been presented in [8], this work employs the hierarchical fuzzy system (HFS) approach, which is effective for tackling the dimensionality problem to build predictive models for negotiation pricing. The performances of three approaches (i.e., standard FS, SFS-SISOM, and HFS) are further compared and discussed from different perspectives, including interpretability, accuracy, generality, computational cost, and applicability. Moreover, a prototype of an intelligent DSS for negotiation pricing is designed and developed with an integration of these three fuzzy approaches. The IT artifact provides substantial potential and flexibility for end users to choose the most suitable model for the existing negotiation pricing problem.
In the general context of DSS, it is worth distinguishing the features of the proposed negotiation pricing DSS. Most extant studies have been devoted to finding, from a set of known feasible decisions, the best decision to fit with the given set of decision criteria or maximizing the known utility functions. Therefore, these studies can be regarded as DSS with complete and certain information, and the dominant approaches in DSS are decision analysis, ranking, and optimization methods. For the negotiation pricing DSS, however, the situation is very different because the best decision is attainable, and it is the highest price a customer is willing to pay. Yet, the actual problem is that the highest willing-to-pay price is unknown. Therefore, this type of decision-making problem requires DSS with uncertain information. As the distinguishing feature here is the information's uncertainty, the dominant approaches in most existing DSS are no longer applicable and a new approach is needed.
This paper is organized as follows: Section 2 reviews the related work of negotiation pricing decision support, and introduces the HFS and its challenges. In Section 3, a systematic approach employing three different types of FS is proposed and presented to provide decision support for intelligent negotiation pricing. The applicability and utility of the proposed approach is demonstrated and tested against three datasets in Section 4, and the derived results are compared and discussed in Section 5. In Section 6, a prototype system is developed and presented to provide a proof-of-concept for the proposed work. The final section concludes this paper and suggests further work directions.
Section snippets
Negotiation pricing decision support
Negotiation is a crucial activity in business, and is a complex, time consuming, and iterative process which might involve intensive information exchange and processing. In most business negotiations, price is the most important attribute. According to utility theory, in a multi-issue negotiation problems, a utility function can be employed to model price, such that multi-criteria can be converted and evaluated by one dimension [13]. Negotiation pricing aims to identify a mutually beneficial
The proposed approach
Fuzzy set theory has become an increasingly prevalent methodology for representing and dealing with uncertain information, and has been successfully applied to many IT contexts, such as control engineering, soft computing, and intelligent DSSs [6], [7]. The merits of utilizing fuzzy sets for representing subjective expertise/knowledge, handling uncertainty, and modeling reasoning processes have been widely discussed and verified [31]. This study proposes a systematic and learning approach based
Experimental data
Three negotiation pricing related datasets, which vary from the number of instances to dimensionalities, were used in this study. A summary of these three datasets are presented in Table 2.
Interpretability and transparency
In regard to the development of DSSs, interpretability/transparency is one of the most crucially desirable features. It is of paramount importance for researchers to further apply this insight to the predicament of negotiation pricing. The proposed FS approach not only enables effective knowledge discovery, but also efficient knowledge representation in the form of linguistic IF-THEN fuzzy rules, which are understandable and interpretable. When the dimension is low, the standard FS has the best
Developing a prototype system
In this section, a prototype system has been developed and presented. The system is a standalone Java application that provides intelligent decision support for a series of negotiation pricing tasks. The system allows pricing managers to load dataset, design, train, and test the fuzzy model, and then predict the negotiation price/discount for a new transaction. The graphic user interface (GUI) consists of five components: menu, toolbar panel, display panel, tab control panel, and status display
Conclusions
This study proposes a systematic and learning approach consisting of three different FSs (i.e., standard FS, SFS-SISOM, and HFS) to provide intelligent decision support for negotiation pricing, in particular under the high-dimensional and uncertain scenarios. The effectiveness and applicability of the proposed approaches are demonstrated by three experimental datasets, varying from dimensionality to data coverage. The main contributions of this work include: 1) Instead of tackling the difficult
Acknowledgments
This work is supported by the National Natural Science Foundation of China (No. 71301133, 71572166, 71671153, 71671149, and 71202059) and the Fundamental Research Funds for theCentral Universities (Project No. 20720161044). The authors are grateful to the reviewers and editor for their invaluable and insightful comments that have helped to improve this work. Thanks also go to all participants who have helped completed the MP3 dataset questionnaire.
Xin Fu is an Associate Professor in the School of Management, Xiamen University, China. She received her PhD degree in Computer Science from Aberystwyth University, UK. Her research interests include decision support systems, fuzzy and qualitative modelling, business intelligence, and predictive toxicology. Her research has been published in journals including Decision Support Systems, Information & Management, IEEE Transactions on Fuzzy Systems, Pattern Recognition, Journal of Cheminformatics,
References (34)
- et al.
Automated negotiation for e-commerce decision making: a goal deliberated agent architecture for multi-strategy selection
Decis. Support. Syst.
(2015) - et al.
Predicting opponent's moves in electronic negotiations using neural networks
Expert Syst. Appl.
(2008) - et al.
Pricing and promotion strategies of an online shop based on customer segmentation and multiple objective decision making
Expert Syst. Appl.
(2011) - et al.
Fuzzy complex numbers and their application for classifiers performance evaluation
Pattern Recogn.
(2011) - et al.
A negotiation-based scheme for service level pricing for wireless access
Comput. Commun.
(2012) - et al.
Universal approximation by hierarchical fuzzy system with constraints on the fuzzy rule
Fuzzy Sets Syst.
(2002) - et al.
Sellers in e-marketplaces: a fuzzy logic based decision support system
Inf. Sci.
(2014) - et al.
Reservation price reporting mechanisms for online negotiations
Decis. Support. Syst.
(2009) - et al.
A hierarchical genetic fuzzy system based on genetic programming for addressing classification with highly imbalanced and borderline data-sets
Knowl.-Based Syst.
(2013) - et al.
Agent based mobile negotiation for personalized pricing of last minute theatre tickets
Expert Syst. Appl.
(2012)
A hierarchical structure of observer-based adaptive fuzzy-neural controller for MIMO systems
Fuzzy Set. Syst.
Automatic price negotiation on the web: an agent-based web application using fuzzy expert system
Expert Syst. Appl.
Order selection and pricing methods using flexible quantity and fuzzy approach for buyer evaluation
Eur. J. Oper. Res.
Hierarchical fuzzy modelling for monitoring depth of anaesthesia
Fuzzy Sets Syst.
A neural network approach to predicting price negotiation outcomes in business-to-business contexts
Expert Syst. Appl.
Towards an interdisciplinary perspective of training intervention for negotiations: developing strategic negotiation support contents
Decis. Support. Syst.
Steering sales reps through cost information: an investigation into the black box of cognitive references and negotiation behavior
Int. J. Res. Mark.
Cited by (23)
Sparse fuzzy classification for profiling online users and relevant user-generated content
2022, Expert Systems with ApplicationsCitation Excerpt :As a result, data uncertainty becomes an under-addressed issue that impacts on the analysis of online activity, which further limits the application of conventional approaches. Fuzzy set theory is a popular tool used to represent and process uncertain and vague information, while preserving the transparency and interpretability of results (Aghaeipoor, & Javidi, 2020; Fu, Zeng, Luo, Wang, Xu et al., 2017). Accordingly, a diverse range of fuzzy-based algorithms have been proposed for the analysis of online pattern, and some of them are detailed in Section 2.
An integrated multi-stage fuzzy inference performance measurement scheme in humanitarian relief operations
2021, International Journal of Disaster Risk ReductionCitation Excerpt :To avoid the rule-explosion problem, the input to each FIS was restricted to two variables (e.g., B1 and B2, F1 and F2, etc.). Assuming that there are 4 inputs, the architecture requires three stages with two inputs and a single output for each stage to calculate a final crisp output [40,43]. Initially, two input parameters are assigned to a FIS, while subsequently, an output from a previous stage forms an input to a subsequent FIS [40,43].
A flexible rule for evidential combination in Dempster–Shafer theory of evidence
2019, Applied Soft Computing JournalCitation Excerpt :Maybe the most interesting one is to axiomatise our combination rule. Another tempting avenue is to apply our combination rule into some applications, such as multiple sensor surveillance system [20], wireless sensor network [45], and automated e-business negotiation [46–49]. Also, since D–S theory can be used to deal with multiple criteria decision making problem [3,50] (i.e., it can describe both quantitative and qualitative criteria under various uncertainties including ignorance and randomness and the combination rule can be used to combine the ranking of each choice with respect to each criterion), it is interesting to extend our method with criteria weights for multi-criteria decision making.
Developing design principles for the digitalisation of purchasing and supply management
2019, Journal of Purchasing and Supply Management
Xin Fu is an Associate Professor in the School of Management, Xiamen University, China. She received her PhD degree in Computer Science from Aberystwyth University, UK. Her research interests include decision support systems, fuzzy and qualitative modelling, business intelligence, and predictive toxicology. Her research has been published in journals including Decision Support Systems, Information & Management, IEEE Transactions on Fuzzy Systems, Pattern Recognition, Journal of Cheminformatics, etc.
Xiao-Jun Zeng received the B.Sc. degree in mathematics and the M.Sc. degree in control theory and operation research from Xiamen University, China, and the Ph.D. degree in computation from the University of Manchester, U.K. He has been with the University of Manchester since 2002, where he is currently a Senior Lecturer with the School of Computer Science. His current research interests include computational intelligence, machine learning and data mining, decision support systems and game theory with the applications to energy demand side management and pricing. In these areas, he has published more than 120 papers in international journals and conference proceedings. His research impact case Intelligent pricing decision support systems has been selected as one of 20 impact cases studies to highlight impact made by UK academic Computer Science Research within UK and world over the period 2008-2013. Dr. Zeng is an Associate Editor of the IEEE Transactions on Fuzzy Systems and a member of the peer review college of the U.K. Engineering and Physical Sciences Research Council. With KSS Ltd, UK, his research won the European Information Society Technologies Award in 1999 and the Microsoft European Retail Application Developer Awards in 2001 and 2003.
Xin (Robert) Luo is an Endowed Regents Professor and Associate Professor of MIS and Information Assurance in the Anderson School of Management at the University of New Mexico, USA. He received his Ph.D. in MIS from Mississippi State University, USA. He has published research papers in leading journals including European Journal of Information Systems, Decision Support Systems, Communications of the ACM, Journal of the AIS, Journal of Strategic Information Systems, Information Systems Journal, Information & Management, and Computers & Security, etc. He is currently serving as an Ad-hoc Associate Editor for MIS Quarterly and an Associate Editor for Decision Sciences, European Journal of Information Systems, Electronic Commerce Research, Journal of Electronic Commerce Research, and International Conference on Information Systems. His research interests center around information assurance, innovative technologies for strategic decision-making, and global IT management.
Di Wang received the Bachelors degree in computer science and technique from Tianjin University, Tianjin, China, in 2001 and the Ph.D. degree in computer engineering form Nanyang Technological University, Singapore, in 2005. Currently Dr Wang is a senior research with EBTIC, Khalifa University. Her primary research focus has been on social media analytics and specifically, short message analysis. Dr. Wang developed a customer service prototype for Etisalat from tweets. She has also developed a traffic analysis engine using tweets for the department of transport in the UK. In addition, she is also working on an application for smart transport as part of MK Smart, a smart city project in the UK. Dr. Wang has about 50 publications and 2 patent inventions under filling.
Di Xu is a Full Professor of management science and Associate Dean of School of Management at Xiamen University, China. He received his PhD from Xiamen University. His research interests are in the areas of information systems and ecommerce.
Qingliang Fan is currently an Assistant Professor of Economics at the Wang Yanan Institute for Studies in Economics and School of Economics, Xiamen University, China. He earned his Ph.D. in North Carolina State University. His research interests are theoretical and applied econometrics, high dimensional econometrics and big data analysis on consumers and investors behaviors. He has publications in both international and domestic journals, such as Journal of Econometrics. He also developed STATA user defined ado files for some of the methods he published.