Year: 2024
Type: conference-paper
Abstract: We have developed a specialized framework for evaluating online controlled experiments in the highly volatile and rapidly growing skill gaming industry. Addressing challenges such as a limited user base, the outsized influence of top-percentile players, and constantly evolving user behavior, the framework streamlines the experimentation process to deliver timely and accurate insights. Key innovations include the Strata-Smart A/B Experiment Analysis Algorithm (SEAA), which integrates a novel variance reduction technique and a method for distinguishing between outliers and genuine high-value players, alongside a User-Friendly Interface that simplifies analysis for non-technical stakeholders. Additionally, the framework mitigates pre-existing bias by identifying historically similar populations, ensuring robust and reliable conclusions. This system enhances data-driven decision-making, providing consistency and integrity in a fast-paced, competitive environment.
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: gold
OpenAlex ID: W4411624671