Research
Research
Gratuities in a Digital Services Marketplace
Kim, Seung Hyun, On Amir, Kenneth C. Wilbur
Tip request messages are marketing communications that can help balance platform customer spending with worker compensation and effort. We report the first large-scale field experiments that exogenously manipulate motivations in tip request messages, collectively treating 276,006 transactions by 88,857 buyers. We test how appeals to two common tipping motivations–injunctive norms (“It’s customary to leave a tip for the seller’s service”) and reciprocity–affect tipping and related behaviors in a global freelance services marketplace. On the web, first exposures to the injunctive norms message increased new buyers’ tipping rate–defined as the proportion of transactions tipped–by 11.8 percentage points (p.p.), and repeat buyers’ tipping rate by 6.8 p.p. In the app, first exposures to the injunctive norms message increased tipping rate by 2.5 p.p. for new buyers and 1.4 p.p. for repeat buyers. Surprisingly, tipping rate increases come without detectable reductions in platform repatronage or spending. The injunctive norms message effect is larger in transactions with higher prices, 5-star ratings and North American buyers. Collectively, the results suggest that tip request messages can motivate customer tipping to better incentivize and compensate gig workers, and that optimization requires careful testing.
Did Early ChatGPT-4 Adopters Change Payments to Other Digital Services? Exploratory Evidence from a Consumer Spending Panel
Kim, Seung Hyun, Daniel McCarthy, Kenneth C. Wilbur
Management wisdom dictates that customers "hire" products and services to do "jobs." Does generative artificial intelligence threaten other digital services' "jobs"? We investigate consumer spending data to understand how early ChatGPT-4 subscribers changed their spending on other digital services. We employ Coarsened Exact Matching to balance early adopters with later adopters based on pre-ChatGPT-4 payment data. Then we use a triple-difference identification strategy to predict adopters' counterfactual payments to other digital services. We find two main patterns. First, early ChatGPT-4 adopters tended to also pay for other artificial intelligence services. Second, the estimates rule out market share declines larger than 2% for nearly all other digital service brands and categories. We interpret these non-findings as null effects, providing early evidence that initial generative artificial intelligence technologies did not seriously undermine other digital services' commercial potential.
Designing Distributed Ledger Technologies, Like Blockchain, for Advertising Markets
Mingyu Joo, Seung Hyun Kim, Anindya Ghose, Kenneth C. Wilbur
Published at IJRM (2023) Link
Distributed Ledger Technologies (DLTs), like Blockchain, could help improve brand safety, consumer privacy and transparency in digital advertising. However, paid advertisements transfer attention, money, and data between three parties: advertiser, consumer, and publisher. Therefore advertising-focused DLTs face more complex design considerations than currency-focused DLTs. We describe four key DLT characteristics: structure, participation/governance, transparency, and terms of exchange. We survey current advertising-focused DLTs and find they each serve only two of the three contracting parties in advertising transactions. We make design recommendations for future advertising-focused DLTs, including a goal of serving both consumers and publishers in addition to advertisers. We also recommend governance, transparency and terms-of-trade considerations. Advertising-focused DLTs have significant promise but also significant obstacles, including multi-sided “chicken-and-egg” problems in standards adoption.