Research

Published & Accepted Papers

• Sophisticated Consumers with Inertia
  with Avner Strulov-Shlain, Navdeep Sahni
  Forthcoming, American Economic Review
  Abstract

  Are consumers aware of their future inertia? We run a field experiment that offers over a million readers of a European newspaper auto-renewing or auto-canceling contracts. Many consumers are inert yet most anticipate and account for their inertia: though offering auto-renewing contracts benefits the firm in the short-term, it lowers subscriptions take-up by 35% and total subscribers by 23% over 20 months. Inertia’s impact on market outcomes depends on consumers’ overall awareness of it, which is often ignored by the literature, firms, and policy makers. In our context, consumer sophistication limits the firm from exploiting their behavioral limitations.

  SSRN
• The Sources of Researcher Variation in Economics
  with Nick Huntington-Klein, Claus Pörtner, and the Many Economists Collaborative (146 research teams)
  Forthcoming, Journal of Economic Literature
  Abstract

  We use a rigorous three-stage many-analysts design to assess how different researcher decisions—specifically data cleaning, research design, and the interpretation of a policy question—affect the variation in estimated treatment effects. A total of 146 research teams each completed the same causal inference task three times each: first with few constraints, then using a shared research design, and finally with pre-cleaned data in addition to a specified design. We find that even when analyzing the same data, teams reach different conclusions. In the first stage, the interquartile range (IQR) of the reported policy effect was 3.1 percentage points, with substantial outliers. Surprisingly, the second stage, which restricted research design choices, exhibited slightly higher IQR (4.0 percentage points), largely attributable to imperfect adherence to the prescribed protocol. By contrast, the final stage, featuring standardized data cleaning, narrowed variation in estimated effects, achieving an IQR of 2.4 percentage points. Reported sample sizes also displayed significant convergence under more restrictive conditions, with the IQR dropping from 295,187 in the first stage to 29,144 in the second, and effectively zero by the third. Our findings underscore the critical importance of data cleaning in shaping applied microeconomic results and highlight avenues for future replication efforts.

  SSRN
• The Impact of the General Data Protection Regulation (GDPR) on Online Tracking
  with Karlo Lukic, Bernd Skiera
  International Journal of Research in Marketing, 43(1), 48-70, 2026
  Abstract

  This study explores the impact of the General Data Protection Regulation (GDPR) on online trackers—vital elements in the online advertising ecosystem. Using a difference-in-differences approach with a balanced panel of 294 publishers, it compares publishers subject to the GDPR with those unaffected (the control group). Drawing on data from WhoTracks.me, which spans 32 months from May 2017 to December 2019, it analyzes how the number of trackers used by publishers changed before and after the GDPR. The findings reveal that although online tracking increased for both groups, the rise was less significant for EU-based publishers subject to the GDPR. Specifically, the GDPR reduced about four trackers per publisher, equating to a 14.79 % decrease compared to the control group. The GDPR was particularly effective in curbing privacy-invasive trackers that collect and share personal data, thereby strengthening user privacy. However, it had a limited impact on advertising trackers and only slightly reduced the presence of analytics trackers.

  SSRN arXiv Data & Code
• Economic Consequences of Online Tracking Restrictions: Evidence from Cookies
  with Bernd Skiera
  International Journal of Research in Marketing, 41(2), 241-264, 2024
  IJRM Best Article Award 2024
  Abstract

  In recent years, European regulators have debated restricting the time an online tracker can track a user to protect consumer privacy better. Despite the significance of these debates, there has been a noticeable absence of any comprehensive cost-benefit analysis. This article fills this gap on the cost side by suggesting an approach to estimate the economic consequences of lifetime restrictions on cookies for publishers. The empirical study on cookies of 54,127 users who received ∼128 million ad impressions over ∼2.5 years yields an average cookie lifetime of 279 days, with an average value of €2.52 per cookie. Only ∼13% of all cookies increase their daily value over time, but their average value is about four times larger than the average value of all cookies. Restricting cookies’ lifetime to one year (two years) could potentially decrease their lifetime value by ∼25% (∼19%), which represents a potential decrease in the value of all cookies of ∼9% (∼5%). Most cookies, however, would not be affected by lifetime restrictions of 12 or 24 months as 72% (85%) of the users delete their cookies within 12 (24) months. In light of the €10.60 billion cookie-based display ad revenue in Europe, such restrictions would endanger €904 million (€576 million) annually, equivalent to €2.08 (€1.33) per EU internet user. The article discusses these results' marketing strategy challenges and opportunities for advertisers and publishers.

  SSRN arXiv
• How Does the Adoption of Ad Blockers Impact News Consumption?
  with Shunyao Yan, Bernd Skiera
  Journal of Marketing Research, 59(5), 1002-1018, 2022
  Abstract

  Ad blockers allow users to browse websites without viewing ads. Online news publishers that rely on advertising income tend to perceive users’ adoption of ad blockers purely as a threat to revenue. Yet, this perception ignores the possibility that avoiding ads—which users presumably dislike—may affect users’ online news consumption behavior in positive ways. Using 3.1 million visits from 79,856 registered users on a news website, this research finds that ad blocker adoption has robust positive effects on the quantity and variety of articles users consume. Specifically, ad blocker adoption increases the number of articles that users read by 21.0%–43.2%, and it increases the number of content categories that users consume by 13.4%–29.1%. These effects are stronger for less-experienced users of the website. The increase in news consumption stems from increases in repeat visits to the news website, rather than in the number of page impressions per visit. These postadoption visits tend to start from direct navigation to the news website, rather than from referral sources. The authors discuss how news publishers could benefit from these findings, including exploring revenue models that consider users’ desire to avoid ads.

  SSRN arXiv
• A De-biased Direct Question Approach to Measuring Consumers' Willingness to Pay
  with Reto Hofstetter, Harley Krohmer, Z. John Zhang
  International Journal of Research in Marketing, 38(1), 70-84, 2021
  Abstract

  Knowledge of consumers' willingness to pay (WTP) is a prerequisite to profitable price-setting. To gauge consumers' WTP, practitioners often rely on a direct single question approach in which consumers are asked to explicitly state their WTP for a product. Despite its popularity among practitioners, this approach has been found to suffer from hypothetical bias. In this paper, we propose a rigorous method that improves the accuracy of the direct single question approach. Specifically, we systematically assess the hypothetical biases associated with the direct single question approach and explore ways to de-bias it. Our results show that by using the de-biasing procedures we propose, we can generate a de-biased direct single question approach that is accurate enough to be useful for managerial decision-making. We validate this approach with two studies in this paper.

  SSRN arXiv
• How Do Consumer Characteristics Affect the Bias in Measuring Willingness to Pay for Innovative Products?
  with Reto Hofstetter, Harley Krohmer, Z. John Zhang
  Journal of Product Innovation Management, 30(5), 1042-53, 2013
  Abstract

  Getting the price right is essential for successful new product introductions. An accurate estimate of consumers' willingness to pay is a crucial part of this task. Measurement of willingness to pay for innovations, however, often yields biased results. In this paper, we investigate consumer‐related characteristics and motives that might underlie this bias. Drawing on the elaboration likelihood model, we develop a conceptual model to identify consumer characteristics relevant for preference measurement for innovative products. In doing so, two main factors that potentially influence hypothetical bias are distinguished: ability and motivation. Our conceptual discussion and empirical results demonstrate that the validity of willingness to pay statements is higher among consumers who show a high ability to assess the new product's utility and who are truly interested in purchasing the new product. Counter to intuition, willingness to pay statements from innovators, consumers with good product category knowledge, or consumers who perceive the new product to be highly innovative are relatively more biased and should be interpreted with caution. This research is among the first to look at consumer characteristics rather than methodological issues when it comes to measuring consumer willingness to pay for innovative products. Our conceptual discussion and empirical examination of the drivers of hypothetical bias can be used to refine the validity of the results of the direct willingness to pay approach. These findings should help improve new product pricing surveys and open new avenues for research in measuring consumer preferences.

  SSRN
• How Should Consumers' Willingness to Pay Be Measured?
  with Reto Hofstetter, Harley Krohmer, Z. John Zhang
  Journal of Marketing Research, 48(1), 172-184, 2011
  Abstract

  This study compares the performance of four commonly used approaches to measure consumers’ willingness to pay with real purchase data (REAL): the open-ended (OE) question format; choice-based conjoint (CBC) analysis; Becker, DeGroot, and Marschak's (BDM) incentive-compatible mechanism; and incentive-aligned choice-based conjoint (ICBC) analysis. With this five-in-one approach, the authors test the relative strengths of the four measurement methods, using REAL as the benchmark, on the basis of statistical criteria and decision-relevant metrics. The results indicate that the BDM and ICBC approaches can pass statistical and decision-oriented tests. The authors find that respondents are more price sensitive in incentive-aligned settings than in non-incentive-aligned settings and the REAL setting. Furthermore, they find a large number of “none” choices under ICBC than under hypothetical conjoint analysis. This study uncovers an intriguing possibility: Even when the OE format and CBC analysis generate hypothetical bias, they may still lead to the right demand curves and right pricing decisions.

  SSRN

Working Papers

• The Economic Value of User Tracking for Publishers
  with Rene Laub, Bernd Skiera
  Revise and Resubmit, Management Science
  Abstract

  Regulators and browsers increasingly restrict user tracking to protect users' privacy online. In two large-scale empirical studies, we study the economic implications for publishers relying on selling advertising space to finance their content. In our first study, we draw on 42 million ad impressions from 111 publishers covering EU desktop browsing traffic in 2016. In our second study, we use 218 million ad impressions from 10,526 publishers (i.e., apps) covering EU and US mobile in-app browsing traffic in 2023. The two studies differ in the share of trackable users (Study 1: 85%; Study 2: Apple: 17%, Android: 91%). Still, we find similar average ad impression price decreases (Study 1: 18% and Study 2: 23%) when user tracking is unavailable. More than 90% of the publishers realize lower prices when selling ad impressions for untrackable users. Publishers offering content on sports, cars, lifestyle & shopping, and news & information suffer the most. Premium publishers with high-quality edited content and strong reputations, thematic-focused (niche) publishers, and smaller publishers suffer less from the unavailability of user tracking. In contrast, non-premium publishers with non-edited or user-generated content, thematic-broad (general news) publishers, and larger publishers suffer more. The availability of a user ID generates the highest value for publishers, whereas collecting a user's browsing history, perceived as intrusive by most users, generates only a small value for publishers. These results affirm that ensuring user privacy online has substantial costs for online publishers, but those costs differ across publishers and the type of collected data. This article offers suggestions to reduce these costs.

  SSRN arXiv
• The Impact of the GDPR on Online Usage Behavior
  with Julia Schmitt, Bernd Skiera
  Revise and Resubmit, Information Systems Research
  Abstract

  Privacy regulations aim to safeguard consumers, but can have unintended consequences on how users interact with websites. This article estimates the causal effect of the EU's General Data Protection Regulation (GDPR) on online usage behavior and decomposes it into usage frequency (unique visitors) and usage intensity (visits per unique visitor). Using a generalized synthetic control estimator across trillions of visits to 6,387 websites in 24 industries and 13 countries-11 months before and 19 months after enforcement-it compares observations subject to the GDPR (EU users or EU websites) with unaffected observations (non-EU users on non-EU websites). Weekly visits decline by 4.88% within 3 months and by 10.02%% after 18 months, with the decline increasingly driven by usage frequency: by 18 months, unique visitors decline by 6.61%, whereas visits per unique visitor decline by only 0.59%. The average conceals offsetting effects consistent with a reallocation of online activity: about one quarter of websites gain significantly, and among websites losing users, the remaining users engage more intensively, whereas intensity falls where user numbers grow. Losses concentrate among hedonic and smaller websites and continue to deepen during the first wave of data-protection enforcement actions rather than being concentrated around the GDPR compliance deadline. Taken together, these patterns are more consistent with restricted data-driven user acquisition and privacy salience, than with consent friction and degraded retention as persistent explanations. Long-horizon, decompositional evaluation thus reveals GDPR's differential effect on usage frequency/intensity and heterogeneous effects not distinguished by the aggregate traffic effects reported in prior work.

  SSRN arXiv
• Towards Developing an Understanding of Consumers' Perceived Privacy Violations in Online Advertising
  with Kinshuk Jerath, Daniel Sokol
  Revise and Resubmit, Production and Operations Management
  Marketing Science Institute (MSI) Working Paper Series, August 2024
  Abstract

  Privacy-enhancing technologies (PETs) represent a critical operational challenge for the online advertising industry, requiring substantial infrastructure investment while promising improved consumer privacy protection. Even when PETs may improve privacy protection from an operational or technical viewpoint, understanding whether PETs actually reduce consumers' perceived privacy violations (PPV) is essential for evaluating their viability. In this research, we characterize advertising practices along the dimensions of tracking and targeting, and understand consumers' PPVs for current practices and proposed PETs through online experiments with U.S. and European consumers. As expected, the industry status quo of behavioral targeting, with high degrees of tracking and targeting, results in high PPV. While new technologies that keep data on users' devices reduce PPV compared to behavioral targeting, the reduction is minimal, including for group-level targeting. Contextual targeting, which involves no tracking, significantly lowers PPV. Not surprisingly, PPV is lowest when tracking is absent, but notably, consumers show similar preferences for untargeted ads and no ads. Importantly, consumer perceptions of privacy violations may not align with technical definitions, suggesting that operational investments in privacy technologies may fail without consumer validation. Therefore, it is essential for managers, industry practitioners, and policymakers to follow a consumer-centric approach to understanding privacy concerns and evaluating the operational viability of privacy-enhancing solutions.

  SSRN arXiv
• Engagement vs. Commitment: The Economic Trade-Offs of Polarizing News Content
  with Shunyao Yan
  Under Review
  Marketing Science Institute (MSI) Working Paper Series, June 2026
  Abstract

  Content that drives engagement need not be the same content that drives willingness to pay. We study how polarizing content affects engagement (time on site) and commitment (subscriptions and retention) on a major news platform. We measure article-level polarization with deep-learning classifiers and large language models tailored to a multiparty system, and identify causal effects with two complementary instrumental variables: a Bartik instrument exploiting supply-side editorial variation, and an election instrument exploiting demand-side political salience. We find that supply-driven increases in polarizing content raise engagement but not subscriptions. During the high-salience election window, the same content reduces subscriptions and accelerates churn, with affective polarization driving the sharpest divergence. On the mechanism, we find evidence inconsistent with confirmation bias: three pre-determined ideology proxies do not moderate the engagement or subscription effects. By contrast, on ideological dimensions where the publisher covers both sides, exogenous shifts in the publisher's supply of content opposite readers' baseline ideology raise their consumption of that content, consistent with balanced consumption. These results document an asymmetric engagement-commitment trade-off for digital publishers: polarizing content reliably captures attention but does not convert to subscriptions, and actively damages commitment when political salience is elevated

  SSRN arXiv
• How Researchers' Analytical Decisions Impact Brand Price Elasticity Estimates: Insights from a Many-Analyst Study
  with Koen Pauwels, Karin Teichmann, Oliver Koll, Hannes Datta, Harald van Heerde, and the IJRM Reproducibility Collaboration
  PsyArXiv Preprint, July 2026
  Abstract

  Price elasticities are central empirical measures in marketing, informing pricing, promotion planning, or evaluations of consumer responsiveness. Yet estimates may depend not only on brands, categories, and markets but also on the analytical decisions researchers make in transforming raw purchase data into elasticities. We use a many-analyst design in which 156 independent research teams (RTs) estimate brand price elasticities for 68 meat-substitute country-brand combinations from the same dataset. Each team retained discretion over its analytical approach, including choices concerning modeling, aggregation, fixed effects, control variables, and endogeneity correction. Results show substantial heterogeneity in both choices and estimates. Across 10,415 brand price elasticity estimates, the average is -0.61. However, the variation across the 156 RTs is considerable: the average brand price elasticity per RT ranges from -7.3 to +3.1, with a standard deviation of 0.84. Choices related to time aggregation, fixed effects, time controls, and marketing-mix covariates explain the largest share of variation in reported estimates. Because analytical choices occur in configurations rather than in isolation, we cluster RTs and reveal distinct analytical pathways associated with different patterns of estimates. Our findings contribute to elasticity research by identifying modeling decisions that warrant particular scrutiny and to the open science debate by supporting the call for greater transparency about analytical decisions.

• Paying for Privacy: Pay-or-Tracking Walls
  with Timo Müller-Tribbensee, Bernd Skiera, Reinhold Kesler
  Work in progress
  Abstract

  Prestigious news publishers, and more recently, Meta, have begun to request that users pay for privacy. Specifically, users receive a notification banner, referred to as a pay-or-tracking wall, that requires them to (i) pay money to avoid being tracked or (ii) consent to being tracked. These walls have invited concerns that privacy might become a luxury. However, little is known about pay-or-tracking walls, which prevents a meaningful discussion about their appropriateness. This paper conducts several empirical studies and finds that top EU publishers use pay-or-tracking walls. Their implementations involve various approaches, including bundling the pay option with advertising-free access or additional content. The price for not being tracked exceeds the advertising revenue that publishers generate from a user who consents to being tracked. Notably, publishers' traffic does not decline when implementing a pay-or-tracking wall and most users consent to being tracked; only a few users pay. In short, pay-or-tracking walls seem to provide the means for expanding the practice of tracking. Publishers profit from pay-or-tracking walls and may observe a revenue increase of 16.4% due to tracking more users than under a cookie consent banner.

  SSRN arXiv

Resting Papers

• How Much Ad Viewability is Enough? The Effect of Display Ad Viewability on Advertising Effectiveness
  with Christina Uhl, Nadia Abou Nabout
  Working paper, 2020
  Abstract

  A large share of all online display advertisements (ads) are never seen by a human. For instance, an ad could appear below the page fold, where a user never scrolls. Yet, an ad is essentially ineffective if it is not at least somewhat viewable. Ad viewability - which refers to the pixel percentage-in-view and the exposure duration of an online display ad - has recently garnered great interest among digital advertisers and publishers. However, we know very little about the impact of ad viewability on advertising effectiveness. We work to close this gap by analyzing a large-scale observational data set with more than 350,000 ad impressions similar to the data sets that are typically available to digital advertisers and publishers. This analysis reveals that longer exposure durations (>10 seconds) and 100% visible pixels do not appear to be optimal in generating view-throughs. The highest view-through rates seem to be generated with relatively lower pixel/second-combinations of 50%/1, 50%/5, 75%/1, and 75%/5. However, this analysis does not account for user behavior that may be correlated with or even drive ad viewability and may therefore result in endogeneity issues. Consequently, we manipulated ad viewability in a randomized online experiment for a major European news website, finding the highest ad recognition rates among relatively higher pixel/second-combinations of 75%/10, 100%/5 and 100%/10. Everything below 75% or 5 seconds performs worse. Yet, we find that it may be sufficient to have either a long exposure duration or high pixel percentage-in-view to reach high advertising effectiveness. Our results provide guidance to advertisers enabling them to establish target viewability rates more appropriately and to publishers who wish to differentiate their viewability products.

  SSRN arXiv

Industry Publications

• Should Your Subscription Business Use Auto-Renew?
  with Z. John Zhang
  Harvard Business Review, May 2026
• Business Leaders Should Learn From Facebook’s Decline
  Forbes, March 2022
• Cookies Are Worth More Than You Think
  with Bernd Skiera
  Harvard Business Review, Nov–Dec 2024 (Print Edition)

Conference Proceedings

• How Does Disabling Cookie Tracking Impact Online News Consumption?
  with Yufei Shen, Xitong Li
  International Conference on Information Systems (ICIS), Bangkok, 2024

Book

• The Impact of the General Data Protection Regulation (GDPR) on the Online Advertising Market
  with Bernd Skiera, Yuxi Jin, Lennart Kraft, Rene Laub, Julia Schmitt
  gdpr-impact.com