ABSTRACT:
Two papers that open this issue of JMIS engage at a deep level the human-machine interactions that permeate our everyday life. The first of these presents a study of this interaction in the context of multi-sided recommenders, the second is in the context of conversational agents. The first of the articles aims to limit the negative network effects that can emerge on platforms; the second work studies the role of the human-like aspects of the chatbots. In summary, both works investigate how information technology (IT) can be deployed more effectively in the service of the markets.
The first paper, by Onkar S. Malgonde, He Zhang, Balaji Padmanabhan, and Moez Limayem, studies how to avoid negative direct network effects: if there are many more users on one side of the platform, congestion may result, leading to the decrease of the number of users and the network effects going in reverse. The solution offered here is a novel design of robust recommenders. Robust multi-sided recommenders serve recommendations to both sides of a two-sided platform with the consideration of the uncertainty of data, such as the changing preferences of the agents on both sides of the platform. The authors engage two practical cases in their formal study of such recommenders. With platforms becoming the more definitively chosen format of many markets, the work shows the generalizable practical solutions that formal approaches can offer in a great variety of contexts. The work is also highly generative, with the mapped out future research leading to a further understanding how to maintain the dynamics of the platform marketplaces while upholding positive network effects.
The mimesis of conversational agents imitating humans may be imperceptible when good and irritating when bad. What does it mean? How do we achieve the salutary effects? Shalini Chandra, Anuragini Shirish, and Shirish C. Srivastava use several research methods to investigate which are the desirable human-like competencies of the automated agents in the cognitive, relational, and emotional aspects of their interactions with the actual humans. The work problematizes the notion of naturalness in the chosen research context, and clearly points to further research of this multifaceted construct that will play a major role as we negotiate our putative transition to multiverses and other embeddings of humans in technologically-created environments.
The role of Chief Information Officer (CIO) has been elevated in the recent decade owing to several major factors, including the digitalization of business, the IT-conditioned and highly dynamic competitive field, the percentage of corporate budgets spent on IT, as well as the security exposures brought on by IT. What sort of background should the CIO have? Should the firm seek out a businessperson or a technologist? While an ambidextrous individual would be ideal, generally that is the dilemma. Here, Rajiv D. Banker, Cecilia (Qian) Feng, and Paul A. Pavlou treat the choice in the terms of the alignment between the CIO background and the firm’s strategy (differentiation versus cost containment). With their econometric study of the stock-market reaction to the staffing choice, the authors identify the contingencies of the proper alignment and the deleterious effects of misalignment.
Two subsequent works of the issue investigate empirically the different aspects of collective knowledge creation on the Internet. Xue (Jane) Tan, Fujie Jin, and Alan R. Dennis study the participants’ willingness to offer contributions in electronic networks of practice. The authors dichotomize the feedback the contributors receive from other participants into appreciation (such as positive reviews or high ratings) and attention (number of views accorded the contributions). The three studies, conducted by the researchers in different local and global environments, show consistently that greater appreciation leads to increased future contributions, while greater attention without appreciations decreases them. With further research, the results may perhaps be generalized in the future to other co-creation environments.
Motivating co-creators of value has always been problematic, no less so with many and various opportunities that have now opened for this activity. In the crowdsourcing of ideas, some of the participants are motivated by their own social values, others by their self. The social values may include the altruistic desire to contribute to the wellbeing of others, willingness to reciprocate, and hewing to the community norms. The self-orientation may be fed by a passion for the task, identity construction, self-expression, learning, or striving for social standing, among others. In the next paper, Bei Yan and Andrea B. Hollingshead investigate empirically how these two categories of individuals respond to the motivators enacted by the idea-generation sites. The study convincingly demonstrates that participants’ response to reward structures depends on the contributors’ social value motivation versus self- orientation. Crowdsourcing organizers should design tasks and rewards to motivate participants with both prosocial and self-oriented disposition. It is to be noted that many contributors are moved by a rich array of motivators, yet—as seen here—one of the two categories prevails.
A nodal issue in IT governance is that of centralization versus decentralization of functions, software, capabilities, and other aspects of corporate computing. Balance is sought, and only sometimes achieved. In the next paper, Magno Queiroz, Paul P. Tallon, and Tim Coltman present the outcome of their investigation of the effects of sharing applications among business units on corporate agility. Known as relatedness, the sharing of the apps promises economies and synergies among the firm’s units, while threatening misalignment among them as their specific needs are not fully served. How does app relatedness affect the organization’s responsiveness to the events challenging its competitive standing, or helps the firm challenge others? Will the mandate to share an app developed to the need of another business unit affect the performance of your unit, with its perceived specialized requirements? The authors’ findings, based on the study of 120 organizations surface a convex (U-shaped) relationship. The agility of the unit is affected negatively by the loss of autonomy with respect to the IT applications, before regaining the ground and tending upwards. There is an additional influence of the environmental uncertainty on the shape of the U-curve. The findings are of the moment, as the costs of the apps grow and as the environmental uncertainty is strongly upon us.
Machine-learning (ML) algorithms have become an integral part of information systems. The performance of these algorithms depends on the availability of large datasets. Oftentimes, the big data available for training many of these algorithms is far from evenly distributed across the subject domain, with data sparsity affecting the effectiveness of the ML-based systems and creating biases with far-ranging consequences. Several compensatory techniques have emerged; their performance is uneven. In the next paper, Murtaza Nasir, Ali Dag, Serhat Simsek, Anton Ivanov, and Asil Oztekin present a novel method, relying on a tunable parameter that can be adjusted to the specific dataset’s idiosyncratic imbalance. The authors show the superiority of the algorithm they have developed over the existing methods in most situation of a partial dataset sparsity. This is a significant contribution to a weighty class of information systems (IS) problems.
In our contemporary business environment, we are witnessing a movement toward servitization, that is, marketizing as a service a product that would be otherwise sold as a good. There are numerous variations of this competitive move; a good example is the well-known Otis Elevator scheme of selling their products at a price close to the cost, along with a profitable long-term service contract. The service contract can be properly priced owing to the company’s massive longitudinal data on their products’ use and maintenance needs. The savings owed to predictive maintenance are grounded in such data. The next paper studies the servitization in which the seller firm retains the ownership of the product, while offering it as per-use service. In this business model, IT plays the crucial role, helping co-create economically the value with the consumers. Big data can be employed to optimize the use of leased cars, as a simple example. Xin Zhang, Xiaolong Guo, Wei T. Yue, and Yugang Yu offer a formal model, comparing the effects of a data-centric business model such as servitization with the more traditional good-selling model. The general results are somewhat troubling. The authors show that the superiority of the servitization model for the seller results only when a firm’s data-based service improvement capability is relatively high. However, that capability may lead to a negative environmental impact. This finding leads the authors to indicate the circumstances under which using the servitization model can yield win-win outcomes in terms of profitability and environmental impact.
Two subsequent papers are grounded in the contexts with the rapidly growing IT deployment: smart homes and healthcare. As more and more IT devices enter the home in the progressing implementation of the Internet of Things, they also bring—along with the obvious benefits—security exposures in the intimate domain of our lives. Alaa Nehme and Joey F. George present an aggregated study in behavioral home security, basing themselves in protection-motivation theory. Setting fear appeals against hope appeals, they contribute to our understanding of the coping mechanisms that are effective in securing our increasingly more sophisticated abodes.
Coping in the environment of e-healthcare is the subject of the work authored by Tracy Ann Sykes and Ruba Aljafari. Clearly, the obviously stressful environment aggravated by the integration and evolution of IT devices into the customary routines of physicians and medical paraprofessionals affects not only them but the care they offer to the patients as well (actually, perhaps not as well). The researchers approach the subject from the friendship perspective and show how such relationship ties can relieve some of the stress experienced by the caregivers. Taken together, the two papers contribute significantly to our understanding of how humans and machines can move together to improve lifestyles that integrate newly emerging technologies into routines and change the routines to salutary effects.