Bloomberg was the most expensive data tool in our stack and the least used. At BCG, I designed and implemented a training system that helped researchers adopt Bloomberg despite severe access constraints. By embedding learning directly into live work, I achieved: - **167% increase** in Bloomberg usage (1.5 → 4 FTEs) - **25 minutes saved** per Bloomberg task - **37% reduction** in client cost on Bloomberg-based work - **Second terminal approved** due to increased demand - **100% positive feedback** across 13 participants over two years Together, these changes turned Bloomberg from a sunk cost into a reliable part of the research workflow. For my work on this project, I received the _Business Builder_ award for operational impact. ### The Problem Bloomberg was underutilized across research teams. Despite its depth and cost, researchers defaulted to cheaper, familiar databases. The goal was simple: **Increase Bloomberg usage without disrupting client work or adding licenses.** I started by interviewing researchers across practice areas to understand barriers. Patterns emerged quickly: - Intimidating interface - Limited understanding of Bloomberg’s strengths - Strong habits built around alternate tools - Friction caused by single-terminal access ### Constraints - Only one Bloomberg terminal in the office - Client work always took priority - Learners needed hands-on practice to build confidence - No room for traditional classroom training ### Design Decisions These constraints shaped the following **key design decisions**: - Replaced scheduled training with task-triggered learning - Prioritized high-frequency workflows over full feature coverage - Redesigned my own work schedule to unblock development under terminal constraints Learners watched short, focused videos (5–8 minutes) mapped directly to real Bloomberg workflows, and learning activated _only when a task required Bloomberg_. After watching a video, they immediately executed the task on the terminal. I reviewed all initial outputs to maintain quality and mitigate delivery risk. This allowed learning to happen without competing with billable work. I prioritized workflows that accounted for ~80% of Bloomberg value. Training materials included: - Task-based video walkthroughs - Comparative guides showing when Bloomberg outperformed other tools - Troubleshooting documentation and escalation paths - Lightweight job aids for on-the-job reference The pilot launched with six researchers across key practice areas. I provided direct feedback on all early work. Over time, trained researchers supported new hires, making the system self-sustaining. The system continues to onboard new researchers without additional licenses or formal training time.