2025- Reprints: Technology and Information Systems
חיזוי ביצועי ציוותי עובדים במשימות טוריות בשווקי העבודה המקוונים, חידושים בניהול, 15, 2024
א. שריקי ות. גבע
(Reprint No.: 447)
Research No.: 00921100
בעבודה זו התמקדנו בבעיות המאתגרות של חיזוי תוצאות שיבוץ עובדים בשוקי העבודה המקוונים. שווקים אלו מאפשרים גיוס רבבות עובדים לתקופת זמן קצרה למשימות מוגדרות. התמקדנו בחיזוי ביצועי צמדי עובדים הפועלים על משימה טורית משותפת בשוקי עבודה מקוונים, כמקרה בסיסי של עבודת צוות בשווקים אלו. באופן ספציפי בחנו את יכולת החיזוי (Predictability) של איכות ביצועי עובדים משותפים המבוססת על עבודת כל עובד בנפרד וגם על הסינרגיה בין העובדים. השערת המחקר היא כי ישנה תרומה ליכולת החיזוי של ביצועי העובדים המשותפים כאשר משתמשים במידע על מאפייני שני העובדים. כלומר, מודל חיזוי המתבסס על וקטור המאפיינים של שני העובדים יאפשר חיזוי טוב יותר ממודלי חיזוי המתבססים על כל אחד מהעובדים בנפרד. על מנת לבחון זאת מימשנו ניסוי בפלטפורמת העבודה המקוונת Amazon Mechanical Turk שבו ציוותנו אלפי עובדים לטובת ניתוח משותף של כתבות פיננסיות. על ידי שימוש הציוותים הללו כמקורות מידע, ניסינו לחזות את ביצועי העובדים על בסיס מאפייניהם האישיים, כפי שהפלטפורמה מאפשרת לסנן למעסיקיה, ובאמצעות שימוש במידע היסטורי על ביצועי העובדים בחלק מהמקרים. הצלחנו להראות כי בהינתן מידע מועיל בנתונים (כמו נתונים היסטוריים על ביצועי העובד) ושימוש באלגוריתם רגרסיה חזק דיו, קיים שיפור מסוים עבור מודל חיזוי ביצועי הצלחת העובדים המתבסס על נתוני שני העובדים לעומת מודל בסיסי (המסתמך על נתוני אחד העובדים) או נאיבי (המתבסס על ממוצע הצלחת כלל העובדים). עם זאת, השערתנו לא נתמכה במקרה שבו אין נתונים היסטוריים לגבי כל עובד (בעיית .(Cold start תרומת העבודה היא בעצם ניסוח בעיית ציוותי העובדים כבעיה עסקית מבוססת נתונים, ובבחינת יכולת החיזוי של ביצועי עובדים משותפים.
A Machine Learning Framework for Assessing Experts’ Decision Quality, Management Science, 71(7), 5696–5721, 2025
W. Dong, M. Saar-Tsechansky and T. Geva
(Reprint No.: 448)
Research No.: 00920100
https://doi.org/10.1287/mnsc.2021.03357
Expert workers make non-trivial decisions with significant implications. Experts’ decision accuracy is, thus, a fundamental aspect of their judgment quality, key to both management and consumers of experts’ services. Yet, in many important settings, transparency in experts’ decision quality is rarely possible because ground truth data for evaluating the experts’ decisions is costly and available only for a limited set of decisions. Furthermore, different experts typically handle exclusive sets of decisions, and thus, prior solutions that rely on the aggregation of multiple experts’ decisions for the same instance are inapplicable. We first formulate the problem of estimating experts’ decision accuracy in this setting and then develop a machine–learning–based framework to address it. Our method effectively leverages both abundant historical data on workers’ past decisions and scarce decision instances with ground truth labels. Using both semi-synthetic data based on publicly available data sets and purposefully compiled data sets on real workers’ decisions, we conduct extensive empirical evaluations of our method’s performance relative to alternatives. The results show that our approach is superior to existing alternatives across diverse settings, including settings that involve different data domains, experts’ qualities, and amounts of ground truth data. To our knowledge, this paper is the first to posit and address the problem of estimating experts’ decision accuracies from historical data with scarce ground truth, and it is the first to offer comprehensive results for this problem set-ting, establishing the performances that can be achieved across settings as well as the state-of-the-art performance on which future work can build.
The Role of Social Cues and Trust in Users’ Private Information Disclosure, MIS Quarterly, 46 (2), 1109-1133, 2022
L. Zalmanson, G. Oestreicher-Singer and Y. Ecker
(Reprint No.: 450)
Research No.: 05020100
Across different domains, websites are incorporating social media features, rendering themselves interactive and community-oriented. This study suggests that these “friendly” websites may indirectly encourage users to disclose private information. To investigate this possibility, we carried out online experiments utilizing a YouTube-like video-browsing platform. This platform provides a realistic and controlled environment in which to study users’ behaviors and perceptions during their first encounter with a website. We show that the presence of social cues on a website (e.g., an environment in which users “like” or rate website content) indirectly affects users’ likelihood of disclosing private information to that website (such as full name, address, and birthdate) by enhancing users’ “social perceptions” of the website (i.e., their perceptions that the website is a place where they can socialize with others). We further show that the presence of social cues is more likely to enhance users’ social perceptions when users are primed to perceive the website as trustworthy, as opposed to untrustworthy (through the presentation of trust cues such as data protection disclaimers). Moreover, we rule out users’ privacy concerns as an alternative mechanism influencing the relationship between social cues and information disclosure. We ground our observations in goal systems and trust theories. Our insights may be beneficial both for managers and for policy makers who seek to safeguard users’ privacy.
Crowdfunding platforms are believed to create a more equal-opportunity environment for fundraising by removing entrance barriers found in traditional entrepreneurial markets. Our work investigates whether in creating more equal opportunities for entrepreneurs to enter the market, crowdfunding platforms also create a more equal distribution of funds across ventures. To this end, we utilized a natural experiment in the form of a policy change on Kickstarter.com that resulted in opening the market to more players. Using platform-level analysis, we show that opening the platform shifted demand toward the head of the distribution: More funds and backers became concentrated in a smaller number of head-offers (the superstar effect). Using individual backer-level analysis, we show that these changes in demand distribution are likely to have resulted from changes in investors’ investment behavior, beyond changes in the composition of the demand side (that is, changes in the types of investors entering the market). Together, our results suggest that efforts to level the playing field in crowdfunding platforms can ultimately result in a less equitable distribution of funds across market participants, and drive investors to focus their investments on a smaller set of campaigns.
Are We There Yet? Analyzing Progress in the Conversion Funnel Using the Diversity of Searched Products, MIS Quarterly, 46 (4), 2015-2054, 2022
A. Goldstein, G. Oestreicher-Singer, O. Barzilay and I. Yahav
(Reprint No.: 452)
Research No.: 05080100
The conversion funnel is a model describing the stages consumers go through in their journey toward a purchase. This journey often lasts several days to weeks and can include multiple visits to a seller’s website. A large body of literature has focused on using observable search patterns to identify consumers’ hidden purchasing stages and to estimate their likelihood of conversion. We propose a novel set of measures to better reveal the consumer’s hidden stage in the funnel. These measures are based on the diversity of the searches that a customer engages in while browsing an e-commerce website, and they include not only the number of different products that are searched for, but also measures that rely on unobserved similarities among products, captured in a product network (in which products are assumed to be “similar” if they are frequently co-searched). We operationalize and evaluate our proposed measures using a large-scale dataset from a medium-sized tourism website used for comparing and booking flights. We estimate a hidden Markov model to show that our proposed diversity measures are associated with progress in the funnel and consumers’ conversion likelihood. Specifically, we show that consumers go through different distinguishable stages (states) in their journey, characterized by different values of our proposed diversity measures. To demonstrate the managerial and business implications of our theory, we show that incorporating search-diversity measures into a baseline prediction model significantly improves the model’s performance in predicting purchase likelihood and churn.
