Olivia Natan

Publications

Integrating Neuro-Psychological Habit Research into Consumer Choice Models, 2025 [Latest WP Version, Journal]

with Ryan Webb,  Jessica Fong, Peter Landry, Julia Levine, Alex Steiny Wellsjo, Asaf Mazar, Clarice Zhao, Phillippa Lally, Sanne de Wit, John O’Doherty, Andrew Ching, Raphael Thomadsen, Matthew Osborne, Mark Bouton, Wendy Wood, Colin Camerer

Forthcoming, IJRM

Choice Frictions in Large Assortments, 2025 [Latest WP version, Journal]

Marketing Science

Organizational Structure and Pricing: Evidence from a Large U.S. Airline , 2024 [Latest WP Version, Journal]

with Ali Hortaçsu, Hayden Parsley, Timothy Schwieg, and Kevin Williams

The Quarterly Journal of Economics

Demand Estimation with Infrequent Purchases and Small Market Sizes, 2023 [Latest WP Version, Journal]

with Ali Hortaçsu, Hayden Parsley, Timothy Schwieg, and Kevin Williams

Quantitative Economics

Previously Titled "Incorporating Search and Sales Information in Demand Estimation"

Working Papers

Consumer Inferences from Product Rankings: The Role of Beliefs in Search Behavior [SSRN]

with Jessica Fong and Ranmit Pantle

Accepted, Management Science

Stated and Revealed Preferences for AI-Generated Content

with Jessica Fong and H. Tai Lam

As AI-generated content proliferates on social media, the rise of terms like "AI slop" suggests viewers are averse to this content. We study whether this stated aversion translates into revealed preferences using a large-scale dataset of YouTube videos and comments, supplemented by an online experiment. We document that both the prevalence and negativity of comments mentioning AI in YouTube videos have risen over time. Exploiting within-video variation in when an AI mention first enters a video's top comments, we find that the appearance of an AI mention reduces views by 10% on average. The effect exists for short and long-form videos and is concentrated in smaller channels. An online experiment, which holds video content fixed and randomly varies AI disclosure at the selection and viewing stages, confirms these effects. Labeling a video as AI reduces selection probability by 23%, and exposure to an AI callout during viewing reduces watch time by 14% and lowers willingness to engage further with the creator's other videos.

Reviving Durable Digital Goods

with Zi Yang Chen and Mingduo Zhao

We investigate how downloadable content (DLC) affects a video game's core performance and producers' quality incentives. Drawing on add-on pricing and product line design theories, we focus on Steam, covering roughly three-quarters of the U.S. PC market. Our dataset combines daily-level data from Steam's public API and Gamalytic, including estimated sales and revenue for over 50,000 titles. We compare games releasing their first DLC against a matched control group with similar genre, release dates, and price ranges. Using a two-way fixed effects model with imputation estimators, we find DLC launches raise base game sales by roughly 100% on average, with pronounced early spikes. DLC releases also coincide with higher daily reviews and concurrent player counts. On the supply side, DLC releases coincide with a roughly 12 percentage point increase in patch announcements, peaking around day nine. These results suggest DLC enables price discrimination across heterogeneous consumers while revitalizing the base product-through renewed visibility (an advertising effect) and commitments to ongoing maintenance (a signaling effect).

Work in Progress

Targeted Bundling

with Walter W Zhang

We study how digital platforms govern pricing and bundling by examining publishers’ behavior on Steam, the leading PC game distribution platform. In this market, publishers cannot directly personalize prices but can engage in second-degree price discrimination through mixed bundling and promotional discounts. To recover demand from rank and gameplay data, we adapt the semi-parametric approach of Bajari, Fox, and Ryan (2008), which allows us to identify consumer preferences when sales quantities are unobserved. With this model, we analyze the joint decisions of the platform and publishers and evaluate counterfactual bundling and pricing regimes. We compare profit-maximizing strategies, fairness-constrained pricing, and consumer-screening bundles, highlighting implications for platform governance, firm revenues, and consumer welfare. Our case study provides general lessons for how digital platforms can design bundling and discounting policies in the absence of direct personalization.

Is Adtech Consolidation Raising Ad Prices?

with Tesary Lin, Zhengrong Gu, and Samuel G Goldberg

The Value of Data and Intermediary Market Power 

with Tesary Lin, Zhengrong Gu, and Samuel G Goldberg