The Impact of Online Buyer-Determined Auction Design on Bidding Behavior and Auction Outcomes in Online Labor Market

Description

Online labor markets are Internet platforms that facilitate Buyer-Determined (BD) auctions, in which buyers can identify and hire service providers who bid to offer their services. Marketplace platforms such as eLance and Freelancer opened new venues for companies to source services globally and have been growing rapidly. Critical to the success of these marketplaces are the use of BD auctions. Despite the growing significance of the new marketplaces and the critical role online BD auctions play therein, few empirical studies have investigated the impact of auction design in the context of online labor market. To fill this gap, this research studies the impact of BD auction design, specifically bid visibility (i.e., open bid BD auction versus sealed bid BD auction) on service providers’ bidding behavior and consequently on outcomes of BD auctions in online labor market. There has been extensive interest in understanding online auction mechanisms in Information Systems and Economics literature. As a new form of online auctions in the emerging online labor market, online BD auctions present nuances that call for additional empirical investigations that extend the discussion in current auction literature to the new context. To achieve this, I leverage on a unique proprietary dataset collected from a leading marketplace to carry out empirical investigations about (1) the impact of bid visibility on service providers’ bidding behavior (bid anchoring, learning and bid revision) and (2) the impact of bid visibility on auction outcomes (bid price/quality, buyer surplus, contract probability and project performance). In the research, I will implement alternative empirical models and estimation methodologies to ensure the robustness of the research findings. Specifically, I will explore natural experiment conditions created by exogenous changes to the price of auction design features and quasi-experiment conditions generated by matched samples. I will also implement an integrated model of bidding behavior and compare the estimation results to reduced form analyses. In later stage of this project, I will seek for collaboration from the site vendor in conducting field experiments. The proposed research intends to make contributions to the information systems literature on online auctions by empirically studying the impact of bid visibility on bidding behavior and auction outcomes in the context of online labor market. The unique observational data enables empirical comparison of user behavior and auction outcomes under different BD auction design. The study will generate import practical insights for both site vendors and participants in online labor market.