Shield of Eden is a browser extension developed as part of the Marketing for Betterment initiative. It is designed to protect users from manipulative marketing practices while enabling a new wave of ethical, evidence-based research in digital environments.

At its core, the extension serves a dual purpose:

• It functions as a personal shield, informing users about marketing optimizations occurring behind the scenes (such as dynamic pricing, nudging, or A/B-tested content), and gives them the ability to limit or undo certain interventions.
• It also acts as a scientific research platform, offering a realistic, user-based experimental setting where researchers can test hypotheses, run A/B tests, and analyze anonymized behavioral data.

This project introduces a new algorithmic approach to evaluate the unintended consequences of firm-side marketing optimizations. Companies routinely use A/B testing to fine-tune interventions—such as free trial durations, pricing schemes, or ad placements—with the goal of maximizing conversion or revenue. However, these optimizations can sometimes create hidden harm for consumers, particularly when short-term engagement metrics are misaligned with long-term outcomes like satisfaction, retention, or product fit.

Our method offers a way to audit these interventions using the same experimental data firms already collect. By combining treatment effect estimation with a custom recursive partitioning algorithm, we identify user segments where firm and consumer outcomes diverge—surfacing segments that benefit from optimization, and those who may be misled or underserved. The goal is to move beyond averages and uncover where and for whom value is being created, or extracted.

This research contributes a practical tool for businesses seeking more sustainable, fair, and consumer-aware marketing practices. It also opens the door to new regulatory and ethical conversations by offering a transparent mechanism for diagnosing harm in data-driven personalization.

This project investigates how firm-optimized marketing tactics—particularly the use of strategic framing in promotions and product descriptions—can unintentionally influence societal trust. Framing is a powerful tool: companies routinely highlight positive features (“up to 50% off”) while downplaying limitations in fine print. While these tactics are effective in the short term, their cumulative effect on consumer beliefs and social behavior remains poorly understood.

Our hypothesis is that repeated exposure to misleading or manipulative business framing not only reduces trust in firms, but may also lead individuals to lower their trust in others. This normative spillover can occur through two channels: first, by updating beliefs about what others are likely to do (“if businesses deceive, maybe people do too”), and second, by shifting internal norms about what is socially acceptable (“maybe it’s okay to bend the truth”). Both mechanisms could erode generalized trust—an essential public good.

We combine insights from behavioral economics, social psychology, and marketing ethics to design experimental studies that test this spillover pathway. The goal is to uncover whether firm-level marketing practices, optimized for engagement, may unintentionally degrade the moral fabric that sustains social cooperation. This research contributes to an emerging dialogue on the societal externalities of business optimization, offering empirical grounding for more accountable and trust-preserving forms of marketing.