No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Impact of Recommender System on Competition Between Personalizing and Non-Personalizing Firms
Abstract
How do recommender systems affect prices and profits of firms under competition? To explore this question, we model the strategic behavior of customers who make repeated purchases at two competing firms: one that provides personalized recommendations and another that does not. When a customer intends to purchase a product, she obtains recommendations from the personalizing firm and uses this recommendation to eventually purchase from one of the firms. The personalizing firm profiles the customer (based on past purchases) to recommend products. Hence, if a customer purchases less frequently from the personalizing firm, the recommendations made to her become less relevant. While considering the impact on the quality of recommendations received, a customer must balance two opposing forces: (1) the lower price charged by the non-personalizing firm, and (2) an additional fit cost incurred when purchasing from the non-personalizing firm and the increased cost due to recommendations of reduced quality in the future. An outcome of the analysis is that the customers should distribute their purchases across both firms to maximize surplus over a planning horizon. Anticipating this response, the firms simultaneously choose prices. We study the sensitivity of the equilibrium prices and profits of the firms with respect to the effectiveness of the recommender system and the profile deterioration rate. We also analyze some interesting variants of the base model in order to study how its key results could be influenced. One of the key takeaways of this research is that the recommender system can influence the price and profit of not only the personalizing firm but also the non-personalizing firm.
... Based on information technology, e-commerce platforms recommend products to consumers (Pathak et al. 2010;Kumar and Hosanagar 2019). The value of online recommender systems is to help customers quickly find the relative products from thousands of products and generate a lift in sales (De et al. 2010;Ghoshal et al. 2015;Yang and Gao 2017). In Alibaba, recommendation is the key factor for revenues (Wang et al. 2018). ...
... Kumar and Hosanagar (2019) also find that the recommendations can create an 11% boost in sales. Although online recommendations have advantages, not all firms choose to recommend due to high costs (Ghoshal et al. 2015). For example, the online firm iomoio.com, ...
... For example, the online firm iomoio.com, one of the best Mp3 sites, does not provide recommendations for the music it sells (Ghoshal et al. 2015). Hence, for the online platforms, whether to recommend or not is a critical issue. ...
This paper investigates the effects of reselling and agency contracts on the platform’s incentive to share information under two scenarios, i.e., the recommendation scenario and non-recommendation scenario. Through analyzing the equilibrium solutions, we derive the following results. First, comparing the recommendation and non-recommendation scenarios, the platform prefers the recommendation scenario under the reselling contract. However, contrary to traditional wisdom, when the recommendation efficiency is relatively more efficient, the platform chooses the non-recommendation scenario in the agency contract; as the recommendation efficiency becomes less efficient, the recommendation scenario is preferred in the agency contract. Second, for the reselling contract, the platform prefers to withhold demand information in the recommendation and non-recommendation scenarios. However, the platform voluntarily shares this information under the agency contract for the two scenarios. Third, the platform’s optimal strategy depends on the commission rate, the platform’s recommendation efficiency, and the level of information accuracy. Intuitively, a lower commission rate induces the platform to prefer the reselling contract in the recommendation and non-recommendation scenarios. However, a higher commission rate does not always make the platform choose the agency contract. The recommendation efficiency and the level of information accuracy affect the platform’s optimal strategy. Furthermore, given the platform’s optimal strategy, there are Pareto optimal regions that make the manufacturer better off. That is, the platform and the manufacturer achieve a win–win situation. Finally, in an extension of the model, this paper shows the platform’s optimal strategy when a hybrid contract is used, which also demonstrates the robustness of the results.
... These systems-simple or sophisticated, as the userspropose in general suggestions based on past user-item interaction-content-based filtering (CB)-and/or user-usercollaborative filtering (CF)-possible similarities [6][7][8][9], or in [5], where a co-publication network is created based on co-author and author-venue relationships. However, the users may be more sophisticated than the systems and vice versa, and many times the efforts put in customization does not guarantee the success for all user segments [10,11]. Companies are looking to maximize their utility-like revenue, profit, cash flow, credibility, market share, etc.-by recommending as much as possible goods or services that match the users' needs. ...
... Matrix B-Ranking of economic consequences (6) displays the rank of each economic consequence in a criterion, r ij means the rank of product i according to criterion j. For instance, C 1 represents the price and there are 4 products (P1, P2, P3, P 4 ) with the following prices (20,10,30,20). Since from the manager's perspective a higher price means better-30 is the best, but 10 is the worst alternative-then r 11 = 2, r 21 = 3, r 31 = 1, and r 41 = 2. Considering the costhere the lower the better-we may have for the same four ...
... Probably this behavior is due to the very severe penalty induced by the computation mechanism. For the equi important approach (O E ) it is a single penalty given by that 1 2 j from Eq. (9), where j is the absolute rank in criterion i, since for the non-equi approach (O NE ) the penalty being doubled by the weight related multiplication factor w j from Eq. (11), whose computation is depicted in (10). In our case, O E is the worst, being affected not only by the first penalty type, but also by the realism lack in pragmatic use. ...
User decision intuition is challenging and complex, even if the user and product are known. Thus, recommending products is a management decision with high degree of incertitude. What if we are facing also the cold-start problem, like new products or visitors? This is a hot topic in recommender systems, tackled in variously, successfully or not. This perspective adds more incertitude to the existing uncertain scenario. Our philosophy is the shift from a user-centric view, hit by uncertainty, to a company-centric one taken in certainty circumstances, later to apply win–win approaches. We propose a multi-criteria algorithm -MRS OZ- for an ecommerce site RS that tackles the cold-start differently. It uses Onicescu method, being adapted according to Zipf’s Law, very popular in internet marketing. The paper opted for an exploratory research based on primary and secondary methods, consisting in literature review, 2-step survey addressed to 110 managers splat in 2 groups, and statistical analyses. The algorithm may substitute the human expertise on the given sample item list and criteria set. This work reveals that Onicescu method is suitable for recommender systems field, but relative inner category rankings and more domain related weight ratios strengthen the algorithm. Onicescu method has a wide applicability, but not for recommender systems. Also, the mixture with Zipf’s Law is completely experimental in research area.
... Such effects may in turn have profit implications considering that popular items sometimes have lower margins [93]. Finally, competition effects [102,175,299] may also be important to consider, since rewarding higher-margin items could push sellers to increase prices [298], thus impacting customers' willingnessto-pay [7], and market demand [20,282]. ...
Many of today's online services provide personalized recommendations to their users. Such recommendations are typically designed to serve certain user needs, e.g., to quickly find relevant content in situations of information overload. Correspondingly, the academic literature in the field largely focuses on the value of recommender systems for the end user. In this context, one underlying assumption is that the improved service that is achieved through the recommendations will in turn positively impact the organization's goals, e.g., in the form of higher customer retention or loyalty. However, in reality, recommender systems can be used to target organizational economic goals more directly by incorporating monetary considerations such as price awareness and profitability aspects into the underlying recommendation models. In this work, we survey the existing literature on what we call Economic Recommender Systems based on a systematic review approach that helped us identify 133 relevant papers. We first categorize existing works along different dimensions and then review the most important technical approaches from the literature. Furthermore, we discuss common methodologies to evaluate such systems and finally outline the limitations of today's research and future directions.
... Such systems have become essential applications in electronic commerce, providing decision-making support in effectively filtering large amounts of information so that users are directed to the items that best meet their needs. Prior research has shown that recommender systems are critical in promoting sales volumes, firm revenue, and customer retention [14,19,38]. ...
Research on recommender systems has noted that the ranking of recommended items may play an important role in the performance of recommendation algorithms. To advance recommender systems research beyond the traditional approach that ranks recommended products in descending, it is crucial to understand the cognitive processes that online consumers experience when they evaluate products in a sequence. Drawing on evaluability theory and the order effects perspective, we formulate a scenario in which two products are presented sequentially and each product has two attributes, one of which can be evaluated independently while the other is difficult to evaluate without comparison. Analyses show that in two out of the three cases examined, presenting the most recommended product in the second place will result in stronger consumer purchase intentions and willingness to pay. Research hypotheses are proposed based on the results of the scenario analyses and are empirically tested through three laboratory experiments. In Study 1, evidence for the hypothesized order effects is found for the settings with randomly assigned product recommendations. In Study 2, the same effects are observed for the settings with personalized recommendations generated by a collaborative filtering algorithm. In Study 3, it is shown that such order effects also exist in terms of the recommendation strength of recommender systems. These findings provide novel insights into the behavioral implications of using recommender systems in e-commerce, shedding light on additional means of improving the design of such systems.
... Park and Han (2013) studied the effect on diversity of sales due to recommender systems. Ghoshal et al. (2015) found recommender systems to influence both price and profit of companies under competition with one firm providing personalization and other without personalization. Lee and Rha (2016) studied customer's response to personalization-privacy paradox and found that customers consider personalization benefits as a gain. ...
A series of global trends have redefined the nature of organizations and patterns of work. Organizations are giving more weightage to employability skills and individual personality traits rather than subject-relevant knowledge or an applicable degree. There is a widespread mismatch between the job and education in the global market which is affecting both efficiency and recruitments of the workforce It has become essential for the stakeholders to understand the skills and traits that can help foster employability in students. Research identifies the need to empirically explore and examine employability skills. By evaluating the market from the stakeholders’ perspective, the study provides an empirically grounded and holistic understanding of skills that employers deem crucial for students to gain, sustain and enhance employability in the Indian retail industry.
... Nontransactional activities like catalog browsing and product discovery are search-oriented and require IT systems with significant analytics capacity. Recommender systems are commonly implemented by e-commerce websites for this purpose (Ghoshal et al., 2015;Jannach & Jugovac, 2019). E-commerce websites are both an information system and a marketing channel, so the design and maintenance of e-commerce websites need to follow a customer-focused approach (Albert et al., 2004). ...
Information technology (IT) is a critical resource and asset for a firm’s internal operations and external interactions with outside stakeholders. This paper explores IT capabilities that enable a retailer’s value chain activities and interfaces with suppliers and customers. IT capabilities that enable the value chain activities of logistics and operations as well as those of marketing and sales are identified to examine how they influence a retailer’s implementation of value chain interfaces. Through the lens of organizational information processing theory (OIPT), we find that both IT capabilities and IT infrastructure play a pivotal role in implementing value chain interfaces but in distinctive ways. On the supplier-facing side, a retailer’s IT infrastructure and IT capabilities enabling logistics and operations are essential to its vendor-managed inventory (VMI) system. On the customer-facing side, a retailer’s e-commerce website requires IT infrastructure and IT capabilities that enable marketing and sales. Moreover, synergies are found downstream for e-commerce website that benefits from the integration of IT infrastructure and IT capabilities enabling logistics and operations while such synergies are not found upstream for VMI. This partial complementary effect is explicated by the sequential asymmetric information dependency of value chain activities. Finally, we discuss the implications of our findings for the research and practice of IT enablement and integration.
... From CL, the algorithm selects N articles for display as "recommendations" at each time index 't'. Generally, recommendations are offered in personalized or non-personalized way [33]. We focus on non-personalized top-N news recommendations. ...
Algorithms are increasingly making decisions regarding what news articles should be shown to online users. In recent times, unhealthy outcomes from these systems have been highlighted including their vulnerability to amplifying small differences and offering less choice to readers. In this paper we present and study a new class of feedback models that exhibit a variety of self-organizing behaviors. In addition to showing important emergent properties, our model generalizes the popular “top-N news recommender systems” in a manner that provides media managers a mechanism to guide the emergent outcomes to mitigate potentially unhealthy outcomes driven by the self-organizing dynamics. We use complex adaptive systems framework to model the popularity evolution of news articles. In particular, we use agent-based simulation to model a reader’s behavior at the microscopic level and study the impact of various simulation hyperparameters on overall emergent phenomena. This simulation exercise enables us to show how the feedback model can be used as an alternative recommender to conventional top-N systems. Finally, we present a design framework for multi-objective evolutionary optimization that enables recommendation systems to co-evolve with the changing online news readership landscape.
... They show that data analytics can transform data into insights by the dynamics processes and technology, providing more value for robust decision-making and business problem solutions for ecommerce companies. Some growth areas of e-commerce research include advertising strategies and recommendation systems for online firms (Ghoshal et al., 2015). Web and mobile advertising have also been other interesting areas of research. ...
The era of big data brings unprecedented opportunities and challenges to management research. As one of the important functions of management decision-making, evaluation has been given more functions and application space. Exploring the applicable evaluation methods in the big data environment has become an important subject of research. The purpose of this paper is to provide an overview and discussion of systematic evaluation and improvement in the big data environment. We first review the evaluation methods based on the main analytic techniques of big data such as data mining, statistical methods, optimization and simulation, and deep learning. Focused on the characteristics of big data (association feature, data loss, data noise, and visualization), the relevant evaluation methods are given. Furthermore, we explore the systematic improvement studies and application fields. Finally, we analyze the new application areas of evaluation methods and give the future directions of evaluation method research in a big data environment from six aspects. We hope our research could provide meaningful insights for subsequent research.
... Although online RS are omnipresent in an increasing number of digital environments and are an important factor for competition and price setting among firms (Ghoshal et al., 2015), there is still conflicting evidence for their effects on the diversity of consumption or purchases. Our goal was to close this research gap since both vendors and product suppliers need to know more about the consequences of RS usage. ...
Purpose
The purpose of this paper is to explore the effects of online recommender systems (RS) on three types of diversity: algorithmic recommendation diversity, perceived recommendation diversity and sales diversity. The analysis distinguishes different recommendation algorithms and shows whether user perceptions match the actual effects of RS on sales.
Design/methodology/approach
An online experiment was conducted using a realistic shop design, various recommendation algorithms and a representative consumer sample to ensure the generalizability of the findings.
Findings
Recommendation algorithms show a differential impact on sales diversity, but only collaborative filtering can lead to higher sales diversity. However, some of these effects are subject to how much information firms have about users’ preferences. The level of recommendation diversity perceived by users does not always reflect the factual diversity effects.
Research limitations/implications
Recommendation and consumption patterns might differ for other types of products; future studies should replicate the study with search or credence goods. The authors also recommend that future research should move from taking a unidimensional measure for the assessment of diversity and employ multidimensional measures instead.
Practical implications
Online shops need to conduct a more comprehensive assessment of their RS’ effect on diversity, taking into account not only the effects on their sales distribution, but also on users’ perceptions and faith in the recommendation algorithm.
Originality/value
This study offers a framework for assessing different forms of diversity in online RS. It employs various recommendation algorithms and compares their impact using not just one but three different types of diversity measures. This helps explaining some of the contradictious findings from the previous literature.
... In particular, prior work investigated the influence of recommender systems on purchase decision processes (Dellaert and Häubl, 2012;Parra and Ruiz, 2009;Pathak et al., 2010). Recommender systems can be considered to be product filtering systems that filter products with mainly historical or implicit consumer input (Ghoshal et al., 2015). These systems help consumers in forming their consideration and choice sets faster and more accurately Häubl and Trifts, 2000). ...
Information has a particular importance in online purchase decision processes. As opposed to consumers in online markets, consumers in online markets cannot inspect the physical product to evaluate it and reduce their perceived risk. Hence, consumers in online markets are dependent upon the information that they can gather about a product in which they are interested. Therefore, they have two primary sources of information: Product descriptions and customer reviews. Both sources affect consumers’ purchase decision processes and hence have an economic impact for consumers, shop providers and manufacturers. It is important to know how these sources of information influence a customer’s online purchase decision and how to extract the relevant information. To examine these research objectives, several studies applying different methodological approaches have been conducted. The results are presented in this dissertation.
... Another important research direction in the retailing domain pertains to the use of recommender systems to improve business outcomes. For example, Ghoshal et al. (2015Ghoshal et al. ( , 2018 utilize game-theoretic models to study the impact of recommender systems on the competition between personalizing and nonpersonalizing firms. They analyze how competing firms can benefit from sharing data with one another. ...
Owing to its multidisciplinary nature, the operations management (OM) and information systems (IS) interface distinguishes itself from the individually focused perspective of both fields. The number and depth of contributions in this department can help both disciplines advance to better address important theoretical and practical challenges of the business world. In this paper, we study the characteristics of problems at the interface between OM and IS, and review past work that has been instrumental in setting the tone and direction of research at this interface. We extend our discussion to provide directions for future research at the OM and IS interface in the domains such as smart city management, healthcare, deep learning and artificial intelligence, fintech and blockchain, Internet of Things and Industry 4.0, and social media and digital platforms.
... Based on their results, they present implementable insights for policy makers regarding how to control wasteful advertising. Ghoshal et al. (2015) find that recommendation systems impact the prices of products in both personalizing and non-personalizing firms. ...
In this day, in the age of big data, consumers leave an easily traceable digital footprint whenever they visit a website online. Firms are interested in capturing the digital footprints of their consumers to understand and predict consumer behavior. In this paper, we study how big data analytics have been used in the domains of information systems, operations management, and healthcare. We also discuss the future potential of big data applications in these domains (especially in the areas of cloud computing, Internet of Things and smart city, predictive manufacturing and 3-D printing, and smart healthcare) and the associated challenges. In this study, we present a framework for applications of big data in these domains with the goal of providing some interesting directions for future research. This article is protected by copyright. All rights reserved.
The success of digital content platforms, such as YouTube, relies on both the creativity of independent content creators and the efficiency of content distribution. By sharing advertising revenue with content creators, these platforms can motivate creators to exert greater effort. Most platforms use recommendation systems to deliver personalized content recommendations to each consumer. As creators’ revenues are contingent on their demand, the demand allocation criteria inherent in the recommendation system can influence their content creation behavior. In this paper, we investigate the influence of a platform’s recommendation system on revenue-sharing plans, content creation, profits, and welfare. Our results show that a platform could benefit by biasing recommendations, that is, recommending content that is not an ideal match to a consumer’s preference, to incentivize creators to produce better-quality content. We refer to this as a biased recommendation strategy. Interestingly, we find that such a biased recommendation strategy may lead to a win-win in which the platform, consumers, and content creators can benefit. Our study also shows that consumers may be worse off when they are more knowledgeable and less dependent on the recommendation system. In addition, the platform, consumers, and creators can benefit when the platform has more accurate information on consumer preferences.
This paper was accepted by Raphael Thomadsen, marketing.
Supplemental Material: The online appendices are available at https://doi.org/10.1287/mnsc.2022.03655 .
We propose a new online-to-offline (O2O) service recommendation method based on a novel customer network and service location (CNLRec) in order to help customer to choose the “ideal” O2O services from a large set of alternatives. Our customer network, based on the “co-used” behaviors obtained from the online rating matrix, captures customers’ online behaviors while service location reflects offline behavior characteristic of the customer. For a target customer, a ranking of candidate services based on their locations and this network is generated, in which customer scale usage bias is eliminated. Our experimental results show that: First, even though the rating matrix is sparse, most customers are connected to our proposed customer network, which largely addresses the problem of sparse data. Second, CNLRec outperforms widely-used and state-of-the-art recommendation methods. In addition, e-commerce recommendations that use CNLRec without including item location information (CNRec) has better performance than existing methods. Third, all attributes in CNLRec, including network attributes (relationship degree and customer attribute) and location attributes, play a significant role in recommendations. Specially, O2O service location plays an important role in O2O service selection. In our research, we find the optimal combinations of these attributes.
Content-based recommendation techniques usually require a large number of training examples for model construction, which however may not always be available in many real-world scenarios. To address the training data availability constraint common to the content-based approach, we develop a collaborative expansion-based approach to expand the size of training examples, which could lead to improved content-based recommendations. We use a book rating data set collected from Amazon to evaluate our proposed method and compare its performance against those of two salient benchmark techniques. The results show that our method outperforms the benchmark techniques consistently and significantly. Our method expands the size of training examples for a focal customer by leveraging the available preferences of his or her referent group, and thereby better supports personalized recommendations than existing techniques that solely follow content-based or collaborative filtering, without incurring costs to identify, collect, and analyze additional information. This study reveals the value and feasibility of collaborative expansion as a viable means to increase training size for the focal customer and thus address the training data availability constraint that seriously hinders the performance of content-based recommender systems.
Should complementary firms offer online personalization services to their customers, and how does the differentiated quality of personalization services affect product prices and profits? To answer these questions, we investigate a two-dimensional model of both vertically differentiated product preferences and horizontally differentiated personalization services. The asymmetric quality of basic and complementary personalization services offered by firms is examined in three cases. The quality asymmetry of basic and complementary personalization services, and the complementarity of products lead to several interesting findings regarding firms’ prices and profits. We find that when differentiated personalization services are offered by firms, the profits for both firms increase in complementarity. Given the presence of complementary personalization services offered by firm 1, both firms are worse off with the quality of complementary personalization services. When quality asymmetry exists for both basic and complementary personalization services, there are win-lose, win-win, and lose-win scenarios, which depend on the level of quality differentiation in the basic personalization services offered by firm 2. Furthermore, by comparing the profits in three cases, we find that firms’ profits rest in complementarity.
As thriving and fast-moving technologies, recommender systems have been widely adopted by online retailers to increase their sales recently. This has significant impacts on the stakeholders in the online supply chain. How an online retailer uses recommender systems to maximise its profit through choosing different recommendation strategies for two upstream competing manufacturers is explored in this paper. In particular, a game between one online retailer and two competitive manufacturers is constructed in which these manufacturers can be selectively and strategically recommended by the retailer. The analytical results show that as the recommendation strength of recommender systems increases, neither manufacturers nor the retailer can always enjoy higher profits, which is counterintuitive. Furthermore, this study reveals that (i) a recommended manufacturer may enjoy a higher profit through sharing the recommendation market with its rival than through monopolising this market; (ii) recommending two manufacturers in both is the most feasible way for the online retailer to benefit from controlling the supply chain. Finally, it is interesting that recommender systems are found to be good mechanisms to help to coordinate the online supply chain with one retailer and two manufacturers because the recommendation market generated by recommender systems alleviates channel conflict.
Collaborative filtering or recommender systems use a database about user
preferences to predict additional topics or products a new user might like. In
this paper we describe several algorithms designed for this task, including
techniques based on correlation coefficients, vector-based similarity
calculations, and statistical Bayesian methods. We compare the predictive
accuracy of the various methods in a set of representative problem domains. We
use two basic classes of evaluation metrics. The first characterizes accuracy
over a set of individual predictions in terms of average absolute deviation.
The second estimates the utility of a ranked list of suggested items. This
metric uses an estimate of the probability that a user will see a
recommendation in an ordered list. Experiments were run for datasets associated
with 3 application areas, 4 experimental protocols, and the 2 evaluation
metrics for the various algorithms. Results indicate that for a wide range of
conditions, Bayesian networks with decision trees at each node and correlation
methods outperform Bayesian-clustering and vector-similarity methods. Between
correlation and Bayesian networks, the preferred method depends on the nature
of the dataset, nature of the application (ranked versus one-by-one
presentation), and the availability of votes with which to make predictions.
Other considerations include the size of database, speed of predictions, and
learning time.
Purpose
In the retailing sector, consumers typically patronize multiple outlets, which leaves outlets striving to earn a greater portion of consumer expenditures. The purpose of this paper is to improve theoretical and empirical knowledge about the impact of retailing loyalty programmes on customer purchasing behaviour.
Design/methodology/approach
The effects of two loyalty programmes on customer behaviour are studied through marketwide panel data on supermarket purchases.
Findings
The impact of loyalty programme membership on customer purchase behaviour is significant.
Research limitations/implications
All behavioural indicators show that members and non‐members of loyalty programmes demonstrate significantly different purchase behaviours, irrespective of other factors. The purchase intensity of cardholders, in terms of total and average shopping baskets, share of purchases, purchase frequency and inter‐purchase time, is significantly higher than that of non‐members throughout the entire three‐year period and the trading areas. The findings require confirmation in other retailing sectors before they may be considered fully generalisable.
Practical implications
Retailers may apply the findings in their attempts to segment their target market, which enables them to allocate their marketing expenditures more effectively.
Originality/value
The study contributes to more “generalisable” knowledge by investigating marketwide scanner panel data about competitive purchasing, loyalty programmes and store locations.
Cooperative (co-op) advertising is an important instrument for aligning manufacturer and retailer decisions in supply chains. In this, the manufacturer announces a co-op advertising policy, i.e., a participation rate that specifies the percentage of the retailer's advertising expenditure that it will provide. In addition, it also announces the wholesale price. In response, the retailer chooses its optimal advertising and pricing policies. We model this supply chain problem as a stochastic Stackelberg differential game whose dynamics follows Sethi's stochastic sales-advertising model. We obtain the condition when offering co-op advertising is optimal. We provide in feedback form the optimal advertising and pricing policies for the manufacturer and the retailer. We contrast the results with the advertising and price decisions of the vertically integrated channel, and suggest a method for coordinating the channel.
While marketing activities increasingly involve personalizing product offers to individually elicited preferences, these unique specifications may not be universally important for product choice. Providing evidence of the limits of treating each customer differently, three experiments show that individuals who exhibit interdependent or collectivistic tendencies tend to be more receptive to recommendations that are not personalized to their own preferences, but instead to the collective preferences of relevant in-groups. However, we find that cultural orientation affects responses to personalized recommendations for only those products whose consumption or choice decision is subject to public scrutiny. We further demonstrate that the favorability of thoughts elicited by ads offering targeted versus personalized offers mediates the effect of cultural orientation on responses to personalization. Finally, both individualistic and collectivistic consumers respond more favorably to offers of targeted recommendations when they believe relevant others share their preferences and when their level of expertise is relatively low.
In this study, we examine firms' incentive to offer customized products in addition to their standard products in a competitive environment. We offer several key insights. First, we delineate market conditions in which firms will (will not) offer customized products in addition to their standard products. Surprisingly, we find that when firms offer customized products they are able to not only expand demand, but can also the prices of their standard products relative to when they do not. Second, we find that when a firm offers customized products it is a dominant strategy for it to also offer its standard product. This result highlights the role of standard products and the importance of retaining them when firms offer customized products. Third, we identify market conditions under which ex ante symmetric firms will adopt symmetric or asymmetric customization strategies. Fourth, we highlight how the degree of customization offered in equilibrium is affected by market parameters. We find that the degree of customization is lower when both firms offer customized products relative to the case when only one firm offers customized products. Finally, we show that customizing products under competition does not lead to a prisoner's dilemma.
This paper investigates the competitive market for mass-customized products. Competition leads to surprising conclusions: Manufacturers customize only one of a product's two attributes, and each manufacturer chooses the same attribute. Customization of both attributes cannot persist in an equilibrium where firms first choose customization and then choose price, because effort to capture market with customization makes a rival desperate, putting downward pressure on prices.
Equilibrium involves partial or no customization. In partial customization, rival firms do not differentiate their mass-customization programs: If firms customize different attributes, many more consumers are indifferent between the two firms. The elasticity of demand is increased and the resulting price war makes differentiated customization unprofitable. If firms customize the same attribute of a two-attribute product, they should concentrate on the attribute with the smaller heterogeneity in consumers' preferences.
We incorporate consumers’ effort in portraying their preferences as a cost of interaction and provide public policy findings on the well-being of these consumers: When this cost is low, consumers are better off with customization than with standard goods, but firms choose too little customization. The loss in consumer surplus is sometimes captured by the firms, but for low interaction costs, firms' profit-driven behavior is economically inefficient.
Sellers who plan to capitalize on the lifetime value of customers need to manage the sales potential from customer referrals proactively. To encourage existing customers to generate referrals, a seller can offer exceptional value to current customers through either excellent quality or a very attractive price. Rewards to customers for referring other customers can also encourage referrals. We investigate when referral rewards should be offered to motivate referrals and derive the optimal combination of reward and price that will lead to the most profitable referrals.
We define a delighted customer as one who obtains a positive level of surplus above a threshold level and, consequently, recommends the product to another customer. We show that the use of referral rewards depends on how demanding consumers are before they are willing to recommend (i.e., on the delight threshold level). The optimal mix of price and referral reward falls into three regions: (1) When customers are easy to delight, the optimal strategy is to lower the price below that of a seller who ignores the referral effect but not to offer rewards. (2) In an intermediate level of customer delight threshold, a seller should use a reward to complement a low-price strategy. As the delight threshold gets higher in this region, price should be higher and the rewards should be raised. (3) When the delight threshold is even higher, the seller should forsake the referral strategy all together. No rewards should be given, and price reverts back to that of a seller who ignores referrals. These results are consistent with the fact that referral rewards are not offered in all markets.
Our analysis highlights the differences between lowering price and offering rewards as tools to motivate referrals. Lowering price is attractive because the seller “kills two birds with one stone”: a lower price increases the probability of an initial purchase and the likelihood of referral. Unfortunately, a low price also creates a “free-riding” problem, because some customers benefit from the low price but do not refer other customers. Free riding becomes more severe with an increasing delight threshold; therefore, motivating referrals through low price is less attractive at high threshold levels. A referral reward helps to alleviate this problem, because of its “pay for performance” incentive (only actual referrals are rewarded.) Unfortunately, rewards can sometimes be given to customers who would have recommended anyway, causing a waste of company resources. The lower the delight threshold level, the bigger the waste and, therefore, motivating referrals through rewards loses attractiveness.
Our theory highlights the advantage of using referral rewards in addition to lowering price to motivate referrals. It explains why referral programs are offered sometimes but not always and provides guidelines to managers on how to set the price and reward optimally.
With personalization, consumers can choose from various product attributes and a customized product is assembled based on their preferences. Marketers often offer personalization on websites. This paper investigates consumer purchase intentions toward personalized products in an online selling situation.The research builds and tests three hypotheses: (1) intention to purchase personalized products will be affected by individualism, uncertainty avoidance, power distance, and masculinity dimensions of a national culture; (2) consumers will be more likely to buy personalized search products than experience products; and (3) intention to buy a personalized product will not be influenced by price premiums up to some level. Results indicate that individualism is the only culture dimension to have a significant effect on purchase intention. Product type and individualism by price interaction also have a significant effect, whereas price does not. Major findings and implications are discussed.
Recent efforts in Web usage mining have started incorporating more semantics into the data in order to obtain a representation
deeper than shallow clicks. In this paper, we review these approaches, and examine the incorporation of simple cues from a
website hierarchy in order to relate clickstream events that would otherwise seem unrelated, and thus perform URL compression.
We study their effect on data reduction and on the quality of the resulting knowledge discovery. Web usage data is also notorious
for containing moderate to high amounts of noise, thus motivating the use of robust knowledge discovery algorithms that can
resist noise and outliers with various degrees of resistance or robustness. Therefore, we also examine the effect of robustness
on the final quality of the knowledge discovery. Our experimental results conclude that post-processed and robust user profiles
have better quality than raw profiles that are estimated through optimization alone. However URL compression, as expected,
tends to reduce the quality, but also can drastically reduce the size of the data set, resulting in faster mining.
Economic modeling provides a formal mechanism to under- stand user incentives and behavior in online systems. In this paper we describe the process of building a parameterized economic model of user contributed ratings in an online movie recommender system. We con- structed a theoretical model to formalize our initial understanding of the system, and collected survey and behavioral data to calibrate an em- pirical model. This model explains 34% of the variation in user rating behavior. We found that while economic modeling in this domain requires an initial understanding of user behavior and access to an uncommonly broad set of user survey and behavioral data, it returns significant formal understanding of the activity being modeled.
We explore the use of social comparison theory as a natural mechanism to increase contributions to an online movie recommendation community by investigating the effects of social information on user behavior in an online field experiment. We find that, after receiving behavioral information about the median user's total number of movie ratings, users below the median demonstrate a 530% increase in the number of monthly movie ratings, while those above the median decrease their monthly ratings by 62%. Movements from both ends converge towards the median, indicating conformity towards a newly-established social norm in a community where such a norm had been absent. Furthermore, the social information has a more dramatic effect on those below the median, suggesting an interaction between conformity and competitive preferences. When given outcome information about the average user's net benefit score from the system, consistent with social preference theory, users with net benefit scores above average contribute 94% of the new updates in the database. In both treatments, we find a highly significant Red Queen Effect. @InProceedings{chen_et_al:DSP:2007:1155, author = {Yan Chen and Maxwell Harper and Joseph Konstan and Sherry Li}, title = {Social Comparisons and Contributions to Online Communities: A Field Experiment on MovieLens}, booktitle = {Computational Social Systems and the Internet}, year = {2007}, editor = {Peter Cramton and Rudolf M{"u}ller and Eva Tardos and Moshe Tennenholtz }, number = {07271}, series = {Dagstuhl Seminar Proceedings}, ISSN = {1862-4405}, publisher = {Internationales Begegnungs- und Forschungszentrum f{"u}r Informatik (IBFI), Schloss Dagstuhl, Germany}, address = {Dagstuhl, Germany}, URL = {http://drops.dagstuhl.de/opus/volltexte/2007/1155}, annote = {Keywords: Social comparison, conformity, public goods, embedded online field experiment} }
Online retailers are increasingly using information technologies to provide value-added services to customers. Prominent examples of these services are online recommender systems and consumer feedback mechanisms, both of which serve to reduce consumer search costs and uncertainty associated with the purchase of unfamiliar products. The central question we address is how recommender systems affect sales. We take into consideration the interaction among recommendations, sales, and price. We then develop a robust empirical model that incorporates the indirect effect of recommendations on sales through retailer pricing, potential simultaneity between sales and recommendations, and a comprehensive measure of the strength of recommendations. Applying the model to a panel data set collected from two online retailers, we found that the strength of recommendations has a positive effect on sales. Moreover, this effect is moderated by the recency effect, where more recently released recommended items positively affect the cross-selling efforts of sellers. We also show that recommender systems help to reinforce the long-tail phenomenon of electronic commerce, and obscure recommendations positively affect cross-selling. We also found a positive effect of recommendations on prices. These results suggest that recommendations not only improve sales but they also provide added flexibility to retailers to adjust their prices. A comparative analysis reveals that recommendations have a higher effect on sales than does consumer feedback. Our empirical results show that providing value-added services, such as digital word of mouth and recommendations, allows retailers to charge higher prices while at the same time increasing demand by providing more information regarding the quality and match of products.
We study a market with customers who have heterogeneous preferences for product attributes. We consider two types of firms that compete on price and product variety: A traditional firm, which chooses a limited set of product configurations, and a customizing firm, which can produce any configuration to order. The traditional firm carries product inventories and experiences a lead-time delay. The customizing firm does not carry inventory, and its customers incur waiting costs until they receive their orders. We assume that the customizing firm has limited capacity in the short run (e.g., when it does not outsource production to high-volume manufacturers). We derive the equilibrium for a duopoly competition between the customizing firm and the traditional firm, study its characteristics, and compare it to a monopoly. We characterize conditions that favor customization under competition. We find that the customizing firm's profit is not monotone in the market size and its ease of customization. Similarly, a decline in the traditional firm's holding cost may increase or decrease its profit. We show that the unit cost differential between the firms crucially affects the customizing firm's ideal market size, its returns from expanding capacity, its product variety, and the way operational improvements affect its performance.
With advances in tracking and database technologies, firms are increasingly able to understand their customers and translate this understanding into products and services that appeal to them. Technologies such as collaborative filtering, data mining, and click-stream analysis enable firms to customize their offerings at the individual level. While there has been a lot of hype about web personalization recently, our understanding of its effectiveness is far from conclusive. Drawing on the elaboration likelihood model (ELM) literature, this research takes the view that the interaction between a firm and its customers is one of communicating a persuasive message to the customers driven by business objectives. In particular, we examine three major elements of a web personalization strategy: level of preference matching, recommendation set size, and sorting cue. These elements can be manipulated by a firm in implementing its personalization strategy. This research also investigates a personal disposition, need for cognition, which plays a role in assessing the effectiveness of web personalization. Research hypotheses are tested using 1,000 subjects in three field experiments based on a ring-tone download website. Our findings indicate the saliency of these variables in different stages of the persuasion process. Theoretical and practical implications of the findings are discussed.
We study optimal pricing in the presence of recommender systems. A recommender system affects the market in two ways: (i) it creates value by reducing product uncertainty for the customers and hence (ii) its recommendations can be offered as add-ons which generate informational externalities. The quality of the recommendation add-on is endogenously determined by sales. We investigate the impact of these factors on the optimal pricing by a seller with a recommender system against a competitive fringe without such a system. If the recommender system is sufficiently effective in reducing uncertainty, then the seller prices otherwise symmetric products differently to have some products experienced more aggressively. Moreover, the seller segments the market so that customers with more inflexible tastes pay higher prices to get better recommendations.
Recommender systems are being used by an ever-increasing number of E-commerce sites to help consumers find products to purchase. What started as a novelty has turned into a serious business tool. Recommender systems use product knowledge -- either hand-coded knowledge provided by experts or "mined" knowledge learned from the behavior of consumers -- to guide consumers through the often-overwhelming task of locating products they will like. In this article we present an explanation of how recommender systems are related to some traditional database analysis techniques. We examine how recommender systems help E-commerce sites increase sales and analyze the recommender systems at six market-leading sites. Based on these examples, we create a taxonomy of recommender systems, including the inputs required from the consumers, the additional knowledge required from the database, the ways the recommendations are presented to consumers, the technologies used to create the recommendations, and t...
In this paper, we present a method to make personalized recommendations when user preferences change over time. Most of the works in the recommender systems literature have been developed under the assumption that user preference has a static pattern. However, this is a strong assumption especially when the user is observed over a long period of time. With the help of a data set on employees 'blog reading behavior, we show that users' product selection behaviors change over time. We propose a hidden Markov model to correctly interpret the users 'product selection behaviors and make personalized recommendations. The user preference is modeled as a hidden Markov sequence. A variable number of product selections of different types by each user in each time period requires a novel observation model. We propose a negative binomial mixture of multinomial to model such observations. This allows us to identify stable global preferences of users and to track individual users through these preferences. We evaluate our model using three real-world data sets with different characteristics. They include data on employee blog reading behavior inside a firm, users 'movie rating behavior at Netflix, and users' music listening behavior collected through last.fm. We compare the recommendation performance ofthe proposed model with that of a number of collaborative filtering algorithms and a recently proposed temporal link prediction algorithm. We find that the proposed HMM-based collaborative filter performs as well as the best among the alternative algorithms when the data is sparse or static. However, it outperforms the existing algorithms when the data is less sparse and the user preference is changing. We further examine the performances ofthe algorithms using simulated data with different characteristics and highlight the scenarios where it is beneficial to use a dynamic model to generate product recommendation.
This paper presents an overview of the field of recommender systems and describes the current generation of recommendation methods that are usually classified into the following three main categories: content-based, collaborative, and hybrid recommendation approaches. This paper also describes various limitations of current recommendation methods and discusses possible extensions that can improve recommendation capabilities and make recommender systems applicable to an even broader range of applications. These extensions include, among others, an improvement of understanding of users and items, incorporation of the contextual information into the recommendation process, support for multicriteria ratings, and a provision of more flexible and less intrusive types of recommendations.
Implicit web usage data is sparse and noisy and cannot be used for usage clustering unless passed through a sophisticated pre-processing phase. In this paper we propose a systematic way to analyze and preprocess the web usage data so that data clustering can be applied effectively to extract similar groups of user. We split the entire process into analysis, preprocessing and outlier detection and show the effect of each phase on Java Application Programming Interface (API) documentation usage data that is collected from our server logs. We use the extracted clusters for web based recommender systems and present the accuracy of the recommendations.
This paper examines the role of consumer preferences, costs, and price competition in determining the competitive product strategy of a firm. In the model studied here, there are two identical firms competing on product quality and price. They face consumers who prefer a higher quality product to a lower quality product, but differ in how much they are willing to pay for quality. The consumers can also choose a substitute if they don't like the product-price offerings of the two firms. For the firms, a higher quality product costs more to produce than a lower quality product. The paper shows that the equilibrium strategy for each firm should be to differentiate its product from its competitor, with the firm choosing the higher quality choosing the higher margin as well. This differentiation, however, is not efficient—that is, it is possible to choose two other products and offer them at prices that cover their marginal cost, and still satisfy consumers' “needs” better in the aggregate. A monopolist, by contrast, would differentiate his product line efficiently. This suggests that cannibalization has different effects on product strategy than competition. The paper also shows that if one firm enters the market first, then it can defend itself from later entrants, and gain a first-mover advantage, by preempting the most desirable product position.
Recommending diverse products to consumers is a new strategy for the next generation of recommender systems. However, no existing studies have empirically identified the impact of product diversity on consumer behavior. The aim of this study is to explain how product category diversity affects customer retention rates. To answer this research question, we examine how the number of product categories purchased by consumers is related to customer retention rates at a large digital content distributor. We use panel data consisting of product characteristics, purchase transactions, and customer retention rates from the company. Through segment-level and individual-level panel data analyses, we find that purchase quantity is positively associated with customer retention rates, and that variety of purchased digital content categories is positively associated with customer retention rates. That is, customers who have purchased digital content from multiple categories are more likely to stay longer than those who purchased digital content from a single category or from fewer categories. Put differently, as a complement to the conventional wisdom that just recommending products with similar features that a customer values highly (i.e., similar content from the same category) is important, our results imply that recommending products with different features (i.e., different content across different categories) is also important.
This book, co-authored by the Nobel-prized economist, Kenneth Arrow, considers public expenditures in the context of modern growth theory. It analyzes optimal growth with public capital. A theory of 'controllability' is developed and injected into public economics and growth models. Originally published in 1970 © The Johns Hopkins University Press 1970, Earthscan 1970, 2011 All rights reserved.
We use a game-theoretic model to examine how information personalization by firms interacts with different dimensions of product differentiation (namely, horizontal and vertical differentiation). We consider the possibility that consumers attach different importance to various types of product differentiation, and report the equilibrium in terms of the "quality-fit" ratio, which measures the relative strength of preference for quality compared to preference for product fit and is a function of the cost of quality and the cost of product misfit. We also consider how different market structures (whether firms are similar or differentiated on the horizontal dimension ex ante) lead to different equilibriums when firms adopt personalization. We show that personalization by one firm leads to higher profits for both firms if product quality and misfit costs are high and the firms offer similar products ex ante. On the other hand, if firms offer differentiated products, personalization is profitable only if the effectiveness of the personalization technology is high or if both product quality and misfit costs are low. We also highlight conditions under which investments in personalization and product quality can be complements or substitutes to each other. Finally, we show that a firm can respond to a competitor's personalization by either increasing (aggressive response) or decreasing (defensive response) investments in its own quality. Our results provide insights to managers on when to invest in personalization technologies and how to adjust their investments in product quality after the firm (or its competitor) adopts personalization.
This paper analyzes interactions between a firm that seeks to discriminate between normal users and hackers that try to penetrate and compromise the firm's information assets. We develop an analytical model in which a variety of factors are balanced to best manage the detection component within information security management. The approach not only considers conventional factors such as detection rate and false-positive rate, but also factors associated with hacker behavior that occur in response to improvements in the detection system made by the firm. Detection can be improved by increasing the system's discrimination ability (i.e., the ability to distinguish between attacks and normal usage) through the application of maintenance effort. The discrimination ability deteriorates over time due to changes in the environment. Also, there is the possibility of sudden shocks that can sharply degrade the discrimination ability. The firm's cost increases as hackers become more knowledgeable by disseminating security knowledge within the hacker population. The problem is solved to reveal the presence of a steady-state solution in which the level of system discrimination ability and maintenance effort are held constant. We find an interesting result where, under certain conditions, hackers do not benefit from disseminating security knowledge among one another. In other situations, we find that hackers benefit because the firm must lower its detection rate in the presence of knowledge dissemination. Other insights into managing detection systems are provided. For example, the presence of security shocks can increase or decrease the optimal discrimination level as compared to the optimal level without shocks.
Decision theory and research have focused almost exclusively on choice—the selection of the best option from a choice set containing two or more options. Largely overlooked is the question of how those particular options got there in the first place—why them and not others? This article describes a theory, called image theory, about how prechoice screening of options governs the contents of the set from which a choice is made and summarizes empirical tests of the theory. The research results suggest that screening plays a far more important role in decision making than is generally appreciated and that our view of decision making must be broadened accordingly.
Despite the proliferation of loyalty programs in a wide range of categories, there is little empirical research that focuses on the measurement of such programs. The key to measuring the influence of loyalty programs is that they operate as dynamic incentive schemes by providing benefits based on cumulative purchasing over time. As such, loyalty programs encourage consumers to shift from myopic or single-period decision making to dynamic or multiple-period decision making. In this study, the author models customers' response to a loyalty program under the assumption that purchases represent the sequential choices of customers who are solving a dynamic optimization problem. The author estimates the theoretical model using a discrete-choice dynamic programming formulation. The author evaluates a specific loyalty program with data from an online merchant that specializes in grocery and drugstore items. Through simulation and policy experiments, it is possible to evaluate and compare the long-term effects of the loyalty program and other marketing instruments (e.g., e-mail coupons, fulfillment rates, shipping fees) on customer retention. Empirical results and policy experiments suggest that the loyalty program under study is successful in increasing annual purchasing for a substantial proportion of customers.
This study examines the behavioral aspect of improving the recommendation agent-consumer relationship, utilizing a model of internal information search for unplanned purchases prompted by a recommendation from a collaborative filtering agent. The model describes how consumers update their beliefs about a product upon receiving a recommendation and identifies the factors affecting the increase in the product's expected utility after the recommendation. A Monte Carlo simulation derives propositions regarding how these factors influence the effectiveness of recommendations. Broadly, the marginal value of recommendation depends on the preference structure of the recipient, the attributes of the product on which the recommendation is based, and the characteristics of the population of consumers. The major managerial implication is that retailers should include more information in recommendations when the products are less common or when there is a large variability of user tastes.
Many firms use decision tools called "automatic recommendation systems" that attempt to analyze a customer's purchase history and identify products the customer may buy if the firm were to bring these products to the customer's attention. Much of the research in the literature today attempts to recommend products that have a high probability of purchase (conditional on the customer's history). However, the author posits that the recommendation decision should be based not on purchase probabilities but rather on the sensitivity of purchase probabilities to the recommendation action. This article attempts to model carefully the role of firms' recommendation actions in modifying customers' buying behaviors relative to what the customers would do without such a recommendation intervention. The author proposes a simple consumer behavior model that accommodates a transparent role for a firm's recommendation actions. The model is expressed in econometric terms so that it can be estimated with available data. The author studies these ideas using purchase data from a real e-commerce firm and compares the performance of the proposed main model with the performance of benchmark models. The author shows that the main model is better than benchmark models on key measures.
Recommender systems assist and augment a natural social process. In a typical recommender system people, provide recommendations as inputs, which tile system then aggregates and directs to appropriate recipients. In some cases, the primary transformation is in the aggregation; in others, the system's value lies in its ability to make good matches between recommenders and those seeking recommendations. This special section includes descriptions of five recommender systems. A sixth article analyzes incentives for provision of recommendations. Recommender systems introduce two interesting incentive problems. First, once one has established a profile of interests, it is easy to free ride by consuming evaluations provided by others. Second, if anyone can provide recommendations, content owners may generate mountains of positive recommendations for their own materials and negative recommendations for their competitors. Recommender systems also raise concerns about personal privacy.
The notion of information transmission occurring in the rewriting process of interactive L systems is made precise. The results concern bounds on the information rate, a synthesis theorem, the undecidability of zero information rate, and the relation between growth rate and information rate.
Since their introduction in the early 1990's, automated recommender systems have revolutionized the marketing and delivery of commerce and content by providing personalized recommendations and predictions over a variety of large and complex product offerings. In this article, we review the key advances in col-laborative filtering recommender systems, focusing on the evolution from research concentrated purely on algorithms to research concentrated on the rich set of ques-tions around the user experience with the recommender. We show through examples that the embedding of the algorithm in the user experience dramatically affects the value to the user of the recommender. We argue that evaluating the user experience of a recommender requires a broader set of measures than have been commonly used, and suggest additional measures that have proven effective. Based on our analysis of the state of the field, we identify the most important open research problems, and outline key challenges slowing the advance of the state of the art, and in some cases limiting the relevance of research to real-world applications.
We present a hidden Markov model for collaborative filtering of implicit ratings when the ratings have been generated by a set of changing user preferences. Most of the works in the collaborative filtering and recommender systems literature have been developed under the assumption that user preference is a static pattern. However, we show by analyzing a dataset on employees’ blog reading behaviors that users’ reading behaviors do change over time. We model the unobserved user preference as a Hidden Markov sequence. The observation that users read variable numbers of blog articles in each time period and choose different types of articles to read, requires a novel observation model. We use a Negative Binomial mixture of Multinomials to model such observations. This allows us to identify stable global preferences of users towards the items in the dataset and allows us to track the users through these preferences. We compare the algorithm with a number of static algorithms and a recently proposed dynamic collaborative filtering algorithm and find that the proposed HMM based collaborative filter outperforms the other algorithms.
We analyze the dynamic strategic interactions between a manufacturer and a retailer in a decentralized distribution channel used to launch an innovative durable product (IDP). The underlying retail demand for the IDP is influenced by word-of-mouth from past adopters and follows a Bass-type diffusion process. The word-of-mouth influence creates a trade-off between immediate and future sales and profits, resulting in a multi-period dynamic supply chain coordination problem. Our analysis shows that while in some environments, the manufacturer is better off with a far-sighted retailer, there are also environments in which the manufacturer is better off with a myopic retailer. We characterize equilibrium dynamic pricing strategies and the resulting sales and profit trajectories. We demonstrate that revenue-sharing contracts can coordinate the IDP's supply chain with both far-sighted and myopic retailers throughout the entire planning horizon and arbitrarily allocate the channel profit.
In several markets, consumers can gain further information regarding how well a product fits their preferences only by experiencing it after purchase. This could then generate loyalty for the products tried first. This paper considers a model in which consumers learn in the first period about the product they buy and then make choices in the second period about the competing products, given what they learned in the first period. The paper finds that if the distribution of valuations for each product is negatively (positively) skewed, a firm benefits (is hurt) in the future from having a greater market share today—the brand loyalty characteristic. With negative skewness, two effects are identified: On one hand, marginal forward-looking consumers are less price sensitive than myopic consumers, and this is a force toward higher prices. On the other hand, forward-looking firms realize that they gain in the future from having a higher market share in the current period and compete more aggressively in prices. For similar discount factors for consumers and firms, the latter effect dominates. The paper also characterizes the importance of consumer learning effects on the market outcome.
As the Netflix Prize competition has demonstrated, matrix factorization models are superior to classic nearest neighbor techniques for producing product recommendations, allowing the incorporation of additional information such as implicit feedback, temporal effects, and confidence levels.
Developing an intelligent recommendation system is a good way to overcome the problem of overloaded products information provided by the e-commerce enterprises. As there are a great number of products on the Internet, it is impossible to recommend all kinds of products in one system. We believe that the personalized recommendation system should be built up according to the special features of a certain sort of product, and forming professional recommendation systems for different products. In this paper, based on the consumer’s current needs obtained from the system-user interactions, we propose a fuzzy-based system for consumer electronics to retrieve optimal products. Experimental results show the system is feasible and effective.
Electronic commerce has enabled the use of intelligent agent technologies that can evaluate buyers, customize products, and price in real-time. Our model of an electronic market with customizable products analyzes the pricing, profitability and welfare implications of agent-based technologies that price dynamically based on product preference information revealed by consumers. We find that in making the trade-off between better prices and better customization, consumers invariably choose less-than-ideal products. Furthermore, this trade-off has a higher impact on buyers on the higher end of the market and causes a transfer of consumer surplus towards buyers with a lower willingness to pay. As buyers adjust their product choices in response to better demand agent technologies, seller revenues decrease since the gains from better buyer information are dominated by the lowering of the total value created from the transactions. We study the strategic and welfare implications of these findings, and discuss managerial and technology development guidelines.
Customer preferences for products are drifting over time. Product perception and popularity are constantly changing as new selection emerges. Similarly, customer inclinations are evolving, leading them to ever redefine their taste. Thus, modeling temporal dynamics should be a key when designing recommender systems or general customer preference models. However, this raises unique challenges. Within the eco-system intersecting multiple products and customers, many different characteristics are shifting simultaneously, while many of them influence each other and often those shifts are delicate and associated with a few data instances. This distinguishes the problem from concept drift explorations, where mostly a single concept is tracked. Classical time-window or instance-decay approaches cannot work, as they lose too much signal when discarding data instances. A more sensitive approach is required, which can make better distinctions between transient effects and long term patterns. The paradigm we offer is creating a model tracking the time changing behavior throughout the life span of the data. This allows us to exploit the relevant components of all data instances, while discarding only what is modeled as being irrelevant. Accordingly, we revamp two leading collaborative filtering recommendation approaches. Evaluation is made on a large movie rating dataset by Netflix. Results are encouraging and better than those previously reported on this dataset.
A number of approaches which use model-based collaborative filtering (CF) for scalability in building recommendation systems in Web personalization have poor accuracy due to the fact that Web usage data is often sparse and noisy. Clustering, mining association rules, and sequence pattern discovery have been used to determine the access behavior model. Making use of some of the characteristics of the modeling process can provide significant improvements to recommendation effectiveness. In an earlier work, we introduced a fuzzy hybrid CF technique which inherits the advantages of both memory-based and model-based CF. In this paper, using relational fuzzy subtractive clustering as the first level modeling and then mining association rules within individual clusters, we propose a two level model-based technique, which is scalable and is an enhancement over association rule based recommender systems. Our results from comprehensive experiments using a large real life Web usage data and performance comparisons with memory-based and model-based approaches help substantiate this claim.
The Internet and related technologies have vastly expanded the variety of products that can be profitably promoted and sold by online retailers. Furthermore, search and recommendation tools reduce consumers’ search costs in the Internet and enable them to extend their search from a few easily found best-selling products (blockbusters) to a large number of less frequently selling items (niches). As a result, Long Tail sales distribution patterns emerge that illustrate an increasing demand in niches. We show in this article how different classes of search and recommendation tools affect the distribution of sales across products, total sales, and consumer surplus. We hereby use an agent-based simulation which is calibrated based on real purchase data of a video-on-demand retailer. We find that a decrease in search costs through improved search technology can either shift demand from blockbusters to niches (search filters and recommendation systems) or from niches to blockbusters (charts and top lists). We break down demand changes into substitution and additional consumption and show that search and recommendation technologies can lead to substantial profit increases for retailers. We also illustrate that decreasing search costs through search and recommendation technologies always lead to an increase in consumer surplus, suggesting that retailers can use these technologies as competitive advantage.
The Internet provides an unprecedented capability for sellers to learn about their customers and offer custom products at special prices. In addition, customization is more feasible today because of advances in manufacturing technologies that have improved sellers' manufacturing flexibility. We first develop a model of product customization and flexible pricing to incorporate the salient roles of the Internet and flexible manufacturing technologies in reducing the costs of designing and producing tailored consumer goods. We show how a monopoly seller may earn the highest profits by producing both standard and custom products and can raise prices for both types of products as customization and information collection technologies improve.
Simultaneous adoption of customization in a duopoly reduces the differentiation between their standard products but does not intensify price competition. Compared with a two-facility monopolist, the duopoly may underinvest in customization. Consumer surplus improves after sellers adopt customization but does not monotonically increase as customization technologies advance. When firms face a fixed entry cost and adopt customization sequentially, the first entrant always achieves an advantage and may be able to deter subsequent entry by choosing its customization scope strategically.
We present a review of research studies that deal with personalization. We synthesize current knowledge about these areas, and identify issues that we envision will be of interest to researchers working in the management sciences. We take an interdisciplinary approach that spans the areas of economics, marketing, information technology, and operations. We present an overarching framework for personalization that allows us to identify key players in the personalization process, as well as, the key stages of personalization. The framework enables us to examine the strategic role of personalization in the interactions between a firm and other key players in the firms value system. We review extant literature on the strategic behavior of firms, and discuss opportunities for analytical and empirical research in this regard. Next, we examine how a firm can learn a customer's preferences, which is one of the key components of the personalization process. We use a utility-based approach to formalize such preference functions, and to understand how these preference functions could be learnt based on a customers interactions' with a firm. We identify well-established techniques in management sciences that can be gainfully employed in future research on personalization.
tion and validationmethods in a system called 1:1Pro. Our approach differsfrom other profiling methods in that we includepersonal behavioral rules in customer profiles.7We can judge the quality of rules stored in customerprofiles in several ways. We might call rules "good"because they are statistically valid, acceptable to ahuman expert in a given application, or effective inthat they result in specific benefits such as better decision-making and recommendation capabilities. ...
Customer preferences for products are drifting over time. Product perception and popularity are constantly changing as new selection emerges. Similarly, customer inclinations are evolving, leading them to ever redefine their taste. Thus, modeling temporal dynamics is essential for designing recommender systems or general customer preference models. However, this raises unique challenges. Within the ecosystem intersecting multiple products and customers, many different characteristics are shifting simultaneously, while many of them influence each other and often those shifts are delicate and associated with a few data instances. This distinguishes the problem from concept drift explorations, where mostly a single concept is tracked. Classical time-window or instance decay approaches cannot work, as they lose too many signals when discarding data instances. A more sensitive approach is required, which can make better distinctions between transient effects and long-term patterns. We show how to model the time changing behavior throughout the life span of the data. Such a model allows us to exploit the relevant components of all data instances, while discarding only what is modeled as being irrelevant. Accordingly, we revamp two leading collaborative filtering recommendation approaches. Evaluation is made on a large movie-rating dataset underlying the Netflix Prize contest. Results are encouraging and better than those previously reported on this dataset. In particular, methods described in this paper play a significant role in the solution that won the Netflix contest.
This paper examines the effect of recommender systems on the diversity of sales. Two anecdotal views exist about such effects. Some believe recommenders help consumers discover new products and thus increase sales diversity. Others believe recommenders only reinforce the popularity of already popular products. This paper is a first attempt to reconcile these seemingly incompatible views. We explore the question in two ways. First, modeling recommender systems analytically allows us to explore their path dependent effects. Second, turning to simulation, we increase the realism of our results by combining choice models with actual implementations of recommender systems. We arrive at four main results. One, some common recommenders lead to a net reduction in average sales diversity. Because common recommenders (e.g., collaborative filters) recommend products based on sales and ratings, they cannot recommend products with limited historical data, even if they would be rated favorably. In turn, these recommenders can create a rich-get-richer effect for popular products and vice-versa for unpopular ones. This finding is often surprising to consumers who express that recommendations have helped them discover new products. In line with this, result two shows it is possible for individual-level diversity to increase but aggregate diversity to decrease; recommenders can push each person to new products, but they often push us toward the same new products. Result three finds that recommenders intensify the effects of chance events on market outcomes. At the product level, recommenders can ‘create hits' out of products with early, high sales due to chance alone. At the market level, in individual sample paths it is possible to observe more diversity, even though on average diversity often decreases. Four, we show how basic design choices affect the outcome. Thus, managers can choose recommender designs that are more consistent with their sales or product assortment strategies.
Internet shopbots compare prices and services levels at competing retailers, creating a laboratory for analysing consumer choice. We analyse 20,268 shopbot consumers who select various books from 33 retailers over 69 days. Although each retailer offers a homogeneous product, we find that brand is an important determinant of consumer choice. The three most heavily branded retailers hold a $1.72 price advantage over more generic retailers in head-to-head price comparisons. In particular, we find that consumers use brand as a proxy for retailer credibility in non-contractible aspects of the product and service bundle, such as shipping reliability. Copyright 2001 by Blackwell Publishing Ltd
Recommendation engines are becoming a critical part of many e-commerce sites. The approach uses complex algorithms to analyze large volumes of data and determine what products that potential consumers might want to buy based on their stated preferences, online shopping choices, and the purchases of people with similar tastes or demographics. Recommendation technology must also be able to reach out to small and medium-size businesses and be robust and cost-effective.
Information transmission and recommender systems Available at http:// www.princeton
- D Ozmen
Ozmen, D. Information transmission and recommender systems. Available at http://
www.princeton.edu/smorris/pdfs/PhD/Ozmen.pdf (accessed September 26, 2013).
Personalization hat trick: Revenue, loyalty and conversion. E-commerce Times Product-line competition vs. proliferation
- J H Lovett
- A K Parlaktürk
Lovett, J. Personalization hat trick: Revenue, loyalty and conversion. E-commerce
Times, February 2, 2007.
29. Mendelson, H., and Parlaktürk, A.K. Product-line competition vs. proliferation.
Management Science, 54, 12 (2008), 2039–2053.