Wharton Interactive A/B Testing Simulation

Educational Product December 2021

Interactive simulationConversion optimizationExplore-exploit learning

Overview

I conceived of, co-designed, developed the core simulation logic for, and helped launch Wharton Interactive’s hands-on A/B Testing Simulation. The simulation is designed to be an interactive learning experience where participants step into the role of e-commerce director and use A/B tests to run a CRO campaign over 12 simulated weeks. Learners design experiments, read real-time dashboards, and earn profit by tuning price, copy, and targeting strategies—all inside a browser-based interface that mimics modern CRO platforms. The module can be used in instructional settings to help build intuition about randomized experiments, customer targeting, effect-size × sample-size trade-offs, and the classic explore–exploit dilemma.

Visit Wharton Interactive
Explore Wharton Interactive's experiential learning products

How the Simulation Came Together

The simulation grew from my original product concept. I initiated the project, developed the proposal, helped secure funding, and wireframed the first version before working with academic administrators, product managers, designers, and engineers to turn it into a full educational application.

I built the core Python simulation engine, which generates traffic and heterogeneous treatment effects so that learners must navigate the tension between statistical evidence and economic outcomes. I also worked with Wharton Interactive on an instructor toolkit that allows faculty to adapt the experience to a 2–4-hour class session or an asynchronous assignment.

What Learners Experience

Participants experience the full cycle of an e-commerce conversion-rate-optimization campaign: framing hypotheses, allocating traffic, interpreting lift metrics, and deciding when the evidence is strong enough to act. Because those choices affect simulated profit, statistical concepts remain connected to their business consequences.

The repeated decisions also build intuition for the explore-exploit tradeoff. Learners feel both the cost of acting on noisy data and the opportunity cost of waiting for additional statistical power, rather than encountering those ideas only as abstractions.

Recognition & Impact

Classroom Adoption

Used by more than 5,000 students across dozens of universities, including MIT, the University of Michigan, UCLA, and Duke.

Media & Events

Featured in Wharton’s Analytics Conference; YouTube talk on “The Art & Science of A/B Testing” has 2.5k+ views.