Columbia Business School Research

How has Wikipedia activity changed for articles with content similar to ChatGPT following its introduction? We estimate the impact using differences-in-differences models, with dissimilar Wikipedia articles as a baseline for comparison, to examine how changes in voluntary knowledge contributions and information-seeking behavior differ by article content. Our analysis reveals that newly created, popular articles whose content overlaps with ChatGPT 3.5 saw a greater decline in editing and viewership after the November 2022 launch of ChatGPT than dissimilar articles did.

In this letter, we summarize our recent work on the welfare impact of recommendation algorithms and propose questions for further study. We model recommendation algorithms as an information structure, which shapes how a third party takes actions that affect the welfare of different individuals in a population. Each recommendation algorithm thus induces a welfare profile, describing the expected payoffs of different individuals when the third party takes actions following the algorithm.

We aim to stimulate discussion on how innovation research within marketing can use a better world (BW) perspective to help innovation become a driver of positive change in the world. In this "Challenging the Boundaries" series paper, we hope to provide purposeful research opportunities for scholars seeking to bridge innovation research with the BW movement. We frame our discussion with four areas of innovation research in marketing that are particularly relevant to BW objectives.

Language is a uniquely human trait at the core of human interactions. The language people use often reflects their personality, intentions and state of mind. With the integration of the Internet and social media into everyday life, much of human communication is documented as written text. These online forms of communication (for example, blogs, reviews, social media posts and emails) provide a window into human behaviour and therefore present abundant research opportunities for behavioural science.

One of the most significant levers available to gaming companies in designing digital games is setting the level of difficulty, which essentially regulates the user’s ability to progress within the game. This aspect is particularly significant in free-to-play (F2P) games, where the paid version often aims to enhance the player’s experience and to facilitate faster progression. In this paper, we leverage a large randomized control trial to assess the effect of dynamically adjusting game difficulty on players’ behavior and game monetization in the context of a popular F2P mobile game.

Non-informational cues, such as facial expressions, can significantly influence judgments and interpersonal impressions. While past research has explored how smiling affects business outcomes in offline or in-store contexts, relatively less is known about how smiling influences consumer choice in e-commerce settings even when there is no face-to-face interaction.

Whether speaking, writing, or thinking, almost everything humans do involves language. But can the semantic structure behind how people express their ideas shed light on their future success? Natural language processing of over 40,000 college application essays finds that students whose writing covers more semantic ground, while moving more slowly (i.e. moving between more semantically similar ideas), end up doing better academically (i.e. have a higher college grade point average). These relationships hold controlling for dozens of other factors (e.g.

Using publicly available data from 299 pre-registered replications from the social sciences, we find that the language used to describe a study can predict its replicability above and beyond a large set of controls related to the paper characteristics, study design and results, author information, and replication effort. To understand why, we analyze the textual differences between replicable and nonreplicable studies.

This paper theoretically and empirically investigates the effects of liquidity regulation on the banking system. We document that the current quantity-based liquidity rule has reduced banks’ liquidity risks. However, the mandated liquidity buffer appears to crowd out bank lending and lead to a migration of liquidity risks to banks that are not subject to liquidity regulation. These findings motivate a model of liquidity regulation with endogenous liquidity premiums and heterogeneous banks.

The contemporary art world is conservatively estimated to be a $65 billion USD market that employs millions of human artists, sellers, and collectors globally. Recent attention paid to AI-made art in prestigious galleries, museums, and popular media has provoked debate around how these statistics will change. Unanswered questions fuel growing anxieties. Are AI-made and human-made art evaluated in the same ways? How will growing exposure to AI-made art impact evaluations of human creativity? Our research uses a psychological lens to explore these questions in the realm of visual art.

This paper studies product ranking mechanisms of a monopolistic online platform in the presence of social learning. The products’ quality is initially unknown, but consumers can sequentially learn it as online reviews accumulate. A salient aspect of our problem is that consumers, who want to purchase a product from a list of items displayed by the platform, incur a search cost while scrolling down the list. In this setting, the social learning dynamics, and hence the demand, is affected by the interplay of two unique features: substitution and ranking effects.

As financial technology improves and data becomes more abundant, do market prices reflect this growing information and allocate capital more efficiently? While a number of recent studies have documented rises in aggregate price efficiency, we show that there are large cross-sectional differences. The previously-documented increases are driven by a rise in the informativeness of large, growth stocks. The informational efficiency of smaller assets' prices or prices of assets with less growth potential are either flat or declining.

Ensuring the privacy of training data is a growing concern since many machine learning models are trained on confidential and potentially sensitive data. Much attention has been devoted to methods for protecting individual privacy during analyses of large datasets. However in many settings, global properties of the dataset may also be sensitive (e.g., mortality rate in a hospital rather than presence of a particular patient in the dataset).

With heightened concerns regarding user privacy, there is a recent movement for empowering consumers with the ability to control how their private data are collected, stored, used and shared. Notably, between 2018 and 2020, the General Data Protection Regulation (GDPR) has been implemented in the European Union (EU), and the California Consumer Privacy Act (CCPA) and the California Privacy Rights Act (CPRA) have been implemented/passed in the state of California in the United States. These regulations address both consumer data security and consumer privacy rights.

We explore the relationship between the volatility of a firm's local environment and its organizational structure. Using micro-level data on managers working for a large retailer, we empirically test and provide support for our theory that a more volatile local environment results in more decentralization only when the need for coordination among sub-units is low. In contrast, more local volatility is associated with more centralization when coordination needs are high.

The composition of natural liquidity has been changing over time. An analysis of intraday volumes for the S&P500 constituent stocks illustrates that (i) volume surprises, i.e., deviations from their respective forecasts, are correlated across stocks, and (ii) this correlation increases during the last few hours of the trading session.

Logos serve a fundamental role as the visual figureheads of brands. Yet, because of the difficulty of using unstructured image data, prior research on logo design has largely been limited to nonquantitative studies. In this work, we explore the interplay between logo design and brand identity creation from a data-driven perspective.

Problem definition: We study how online platforms can leverage the behavioral considerations of their users to improve their assortment decisions. Motivated by our collaboration with a dating company, we study how a platform should select the assortments to show to each user in each period to maximize the expected number of matches in a time horizon, considering that a match is formed if two users like each other, possibly on different periods.

Smartphones have made it nearly effortless to share images of branded experiences. This research classifies social media brand imagery and studies user response. Aside from packshots (standalone product images), two types of brand-related selfie images appear online: consumer selfies (featuring brands and consumers’ faces) and an emerging phenomenon the authors term “brand selfies” (invisible consumers holding a branded product).

In modern equity markets, participants have a choice of many exchanges at which to trade. Exchanges typically operate as electronic limit order books under a "price-time" priority rule and, in turn, can be modeled as multi-class FIFO queueing systems. A market with multiple exchanges can be thought of as a decentralized, parallel queueing system. Heterogeneous traders that submit limit orders select the exchange, i.e., the queue, in which to place their orders by trading off financial considerations against anticipated delays until their orders may fill.

We develop a flexible content-based search model that links the content preferences of search engine users to query search volume and click-through rates, while allowing content preferences to vary systematically based on the context of a search. Content preferences are defined over latent topics that describe the content of search queries and search result descriptions.

We evaluate the impacts of tax policy on asset returns using the U.S. municipal bond market. In theory, tax-induced ownership segmentation limits risk-sharing, creating downward-sloping regions of the aggregate demand curve for the asset. In the data, cross-state variation in tax privilege policies predicts differences in in-state ownership of local municipal bonds; the policies create incentives for concentrated local ownership. High tax privilege states have muni-bond yields that are more sensitive to variations in supply and local idiosyncratic risk.

Narratives, and other forms of discourse, are powerful vehicles for informing, entertaining, and making sense of the world. But while everyday language often describes discourse as moving quickly or slowly, covering a lot of ground, or going in circles, little work has actually quantified such movements or examined whether they are beneficial.

Current accounting practice expenses many investments in intangible assets to the income statement, confusing earnings from current revenues with investments to gain future revenues. This has led to increasing calls to book those investments to the balance sheet. Drawing on the relevant research, this paper proposes solutions for the accounting for intangible assets that contrast with balance sheet recognition, and compares them to current practice and the IFRS standards that dictate practice.

Using US Census employer-employee matched data, I show that employer financial distress accelerates the exit of employees to found start-ups. This effect is particularly evident when distressed firms are less able to enforce contracts restricting employee mobility into competing firms. Entrepreneurs exiting financially distressed employers earn higher wages prior to the exit and after founding start-ups, compared to entrepreneurs exiting non-distressed firms. Consistent with distressed firms losing higher-quality workers, their start-ups have higher average employment and payroll growth.

New technologies such as Internet of Things (IoT), Augmented Reality (AR), Virtual Reality (VR), Mixed Reality (MR), virtual assistants, chatbots, and robots, which are typically powered by Artificial Intelligence (AI), are dramatically transforming the customer experience. In this paper, we offer a fresh typology of new technologies powered by AI and propose a new framework for understanding the role of new technologies on the customer/shopper journey.

How do ethical arguments affect AI adoption in business? We randomly expose business decision-makers to arguments used in AI fairness activism. Arguments emphasizing the inescapability of algorithmic bias lead managers to abandon AI for manual review by humans and report greater expectations about lawsuits and negative PR. These effects persist even when AI lowers gender and racial disparities, and when engineering investments to address AI fairness are feasible. Emphasis on status quo comparisons yields opposite effects.

Once artificial intelligence (AI) is indistinguishable from human intelligence, and robots are highly similar in appearance and behavior to humans, there should be no reason to treat AI and robots differently from humans. However, even perfect AI and robots may still be subject to a bias (referred to as speciesism in this article), which will disadvantage them and be a barrier to their commercial adoption as chatbots, decision and recommendation systems, and staff in retail and service settings.

Words are part of almost every marketplace interaction. Online reviews, customer service calls, press releases, marketing communications, and other interactions create a wealth of textual data. But how can marketers best use such data? This article provides an overview of automated textual analysis and details how it can be used to generate marketing insights. The authors discuss how text reflects qualities of the text producer (and the context in which the text was produced) and impacts the audience or text recipient.

Human enhancement products allow consumers to radically enhance their mental abilities. Focusing on cognitive enhancements, we introduce and study a novel factor dehumanization (i.e., denying a person emotional ability and likening them to a robot) which plays a key role in consumers' reluctance to use enhancement products. In study 1, consumers who enhance their mental abilities beyond normal levels were dehumanized, whereas consumers who use the same products to restore lost abilities were not.

Consumer robots are predicted to be employed in a variety of customer-facing situations. As these robots are designed to look and behave like humans, consumers attribute human traits to them—a phenomenon known as the “Eliza Effect.” In four experiments, we show that the anthropomorphism of a consumer robot increases psychological warmth but decreases attitudes, due to uncanniness. Competence judgments are much less affected and not subject to a decrease in attitudes.

This paper modifies the standard returns-earnings regression in accounting research to show that financial reports convey both cash-flow news and discount-rate (expected-return) news. The paper points to the realization principle, associated as it is with the resolution of risk, as the accounting feature that conveys expected-return news. The modified returns-earnings regressions indicate that the information so conveyed pertains to priced risk.

We use five years of bidding data to examine the reaction of advertisers to widely disseminated press on the lack of effectiveness of brand search advertising (queries that contain the firm's name) found in a large experiment run by eBay (Blake, Nosko and Tadelis, 2015). We estimate that 11% of firms that did not face competing ads on their brand keywords, matching the case of eBay, discontinued the practice of brand search advertising.

As a result of the digital revolution, new topics and themes have entered consumer research, and, as the digital revolution enters a new phase, additional new concepts and research questions will emerge. To illustrate the variety of themes on digital technology that consumer researchers have studied, I am presenting a collection of five articles that represent this active new research area. Moreover, I will look into the future and propose a research agenda to address key consumer behavior issues occurring during the next phase of the digital transformation.

Mass customization interfaces typically guide consumers through the configuration process in a sequential manner, focusing on one product attribute after the other. What if this standardized customization experience were personalized for consumers on the basis of how they process information? A series of large-scale field and experimental studies, conducted with Western and Eastern consumers, shows that matching the interface to consumers’ culture-specific processing style enhances the effectiveness of mass customization.

Two modern economic trends are the increase in firm size and advances in information technology. We explore the hypothesis that big data disproportionately benefits big firms. Because they have more economic activity and a longer firm history, large firms have produced more data. As processor speed rises, abundant data attracts more financial analysis. Data analysis improves investors’ forecasts and reduces equity uncertainty, reducing the firm’s cost of capital. When investors can process more data, large firm investment costs fall by more, enabling large firms to grow larger.

On search keywords with trademarked terms, the brand owner ("focal brand") and other relevant firms compete for consumers. For the focal brand, paid clicks have a direct substitute in the organic links below the paid ad(s). The proximity of this substitute depends on whether competing firms are bidding aggressively to siphon off traffic. We study the returns to focal brands and competitors using large-scale experiments on Bing with data from thousands of brands.

In this paper we show how simple text mining and semantic network analysis may be used to (i) improve our theoretical understanding of idea generation, (ii) help people improve the creativity of their ideas. From a theoretical perspective, we contribute to the cognitive idea generation literature by establishing a link between the set of concepts used to form an idea and the creativity of the idea. Each idea contains a subset of the semantic network of concepts related to the topic.

Despite the popularity of prediction, markets among economists, businesses, and policymakers have been slow to adopt them in decision-making. Most studies of prediction markets outside the lab are from public markets with large trading populations. Corporate prediction markets face additional issues, such as thinness, weak incentives, limited entry, and the potential for traders with biases or ulterior motives — raising questions about how well these markets will perform.

We present an end-to-end text mining methodology for relation extraction of adverse drug reactions (ADRs) from medical forums on the Web. Our methodology is novel in that it combines three major characteristics: (i) an underlying concept of using a head-driven phrase structure grammar (HPSG) based parser; (ii) domain-specific relation patterns, the acquisition of which is done primarily using unsupervised methods applied to a large, unlabeled text corpus; and (iii) automated post-processing algorithms for enhancing the set of extracted relations.

I view the research articles presented here as prototypical examples of what may be called “the new wave” in studying Asian markets and consumers. This emerging “new wave” has a different focus than research done over the last few decades. Research is shifting from an emphasis on traditional Asian culture toward a focus on consumer culture and how this consumer culture manifests itself in various Asian markets. The “new wave” research also focuses less on general concepts and more on uniquely Asian phenomena. Finally, methodologically research is shifting from “East” vs.

We report the results of a survey conducted in November 2014 in which 29 quantitative marketing scholars from around the world reflected on the present and future of their field. The survey focused on substantive areas, methods and tools, practical and managerial relevance, doctoral training, and promotion and tenure. The results of the survey revealed several general insights on the challenges and opportunities faced by the field of quantitative marketing research.

The authors discuss the current state and future scenarios of brand experience — a new concept that they contributed to the brand management literature. Specifically, they present three research and practical trends, and marketing challenges: (i) the proliferation of settings and media that evoke brand experiences; (ii) the role of brands in consumption experiences; and (iii) the need of brand experiences to reach positive psychological outcomes.

The value of a business depends on its future not its past. Nonetheless, some managers base key decisions on backwards-looking metrics or models that have been shown to be inappropriate in many situations.

What metrics and models can provide managers with steering control for maximizing the value of their business? This editorial provides some ideas.

We present a method that dynamically designs elicitation questions for estimating preferences, focusing on the parameters of cumulative prospect theory and time discounting models. Typically these parameters are elicited by presenting decision makers with a series of choices between alternatives, gambles or delayed payments. The method dynamically (i.e., adaptively) designs such choices to optimize the information provided by each choice, while leveraging the distribution of the parameters across decision makers (heterogeneity) and capturing response error.