(322b) Data Driven Performance Measurement & Management in the External API Network

Over the past two and a half years, Bristol Myers Squibb (BMS) has shifted from internal scale up and production of clinical small molecule Drug Substance, to external supply from contract manufacturing organizations (CMO`s). This paradigm shift has created a need for BMS to develop new strategies to measure, understand, and manage the performance of our technology transfer & scale up. Our team has developed a culture of performance data collection, gathered across our network of vendors for each project, to better understand performance metrics, such as yield variability, quality variability, key cost drivers, and supply delays. The aggregated, contextualized data enables the creation of fit-for-purpose analysis and reports to support future site selection decisions, vendor management, and internal continuous improvement efforts. Predictive models for business drivers such as project cost and vendor capacity can also be created from the aggregated data. The predictive models can then be used to drive tactical and strategic decisions across the vendor network and project portfolio. These data-driven products are used for multiple business purposes, including high-level performance metrics for API delivery, targeted feedback to vendors, budget estimation, and to support individual strategic decisions around network capacity and capabilities. This presentation will show examples of the workflow used to capture this data, build data-analysis products, and drive decisions and continuous improvement. In particular, we will describe how we measure vendor performance against various business metrics and provide data-driven feedback to partners. We will also cover the use of probabilistic models and Monte-Carlo simulation to understand future expected demand on our supplier network.