Scalability And Business Outcomes: Essays on Managing Trade-Offs When Fringe Technologies Go Mainstream

Abhishek Ray, Purdue University

Abstract

This dissertation consists of three essays that study problems that decision-makers face when hitherto niche technologies scale up. Typically, scaling up involves market expansion with participation from a variety of agents with complex preferences, using the technology to maximize their utility. A major problem for the decision maker then is either one or a combination of the following: deciding policy for optimal business or social outcomes, implementing efficient demand allocation mechanisms or improving market design. The first essay studies policy implications in the context of burgeoning ad-blocking technology. This research problem is important since increased usage of ad-blockers has caused an estimated loss of $21.8 billion in global ad revenues since 2015 and is projected cause $35 billion loss by 2020. The objective of this research is to inform industry and policy-makers on the following questions: (1) what should be the dominant business model of ad-blocking? (2) what are the implications when no such dominant model exists and ad-blockers compete? We consider answering these questions from perspective of three agents: users, a content provider and an ad-blocker. For the first question, we model a monopoly with one ad-blocker and a content provider serving a mass 1 of users. Our model shows under appropriate assumptions, users have a clear preference ordering wherein freemium is the most and paid is the least preferred business model. However content providers and ad-blockers have conflicting preferences on business models, largely dependent on trade-off between advertising revenue and charging users for blocking ads. Further, competition among ad-blockers with same or different business models makes users worse-off and ad-blockers better off, in every case. More generally, from an outcome perspective, our analysis shows that it is impossible to have a dominant ad-blocking business model that satisfies both users, content providers and ad-blockers. The second essay studies optimizing allocations in the context of the growing application of iterative combinatorial auctions in industry. Combinatorial auctions (CA) are increasingly being used to allocate bundles of items among interested bidders. However, with CA being conducted iteratively to ease preference elicitation problems, the volume, variety and velocity of bundles and bids has become complex. It therefore takes exponentially longer to solve for winners in such auctions. As expected, regular solvers such as IBM CPLEX or AMPL have been demonstrably ineffective. We propose an Ant Colony-based algorithm (TrACA) that produces optimal or near-optimal results within specified time. Experiments are performed on 94 instances to show and compare the performance of TrACA to the current state-ofthe-art memetic algorithm (MA) & ant colony based heuristic (ACLS) and a recent exact algorithm.

Degree

Ph.D.

Advisors

Nguyen, Purdue University.

Subject Area

Design|Business administration|Information Technology|Marketing|Mass communications

COinS