OBJECTIVES: Big data has the potential to provide tremendous value to healthcare manufacturers, improving their understanding of unmet clinical need and informing product development. The objective of this study was to analyze leading payer claims databases and EMR systems for diagnostic specific information and costs and to determine where unmet needs and opportunities for future data optimization exist.   METHODS: Five companies who sell large claims and EMR data sets were interviewed to understand costs and data granularity.  Additionally, interviews were conducted with three diagnostic companies that recently purchased big data sets to better understand the opportunities and limitations of the data purchased related to diagnostic decision-making and research.    RESULTS: Datasets reviewed contained little granularity related to diagnostic tests. Specifically, because individual tests cannot be determined by CPT code, there was no way to determine the brand of test used or whether tests were FDA-approved or laboratory developed.  Additionally, neither claims databases nor EMR systems capture the diagnostic platform used for laboratory analysis.  There was variation in the detail contained in the databases related to lab results.  EMR systems seemed to contain greater detail than claims system, but lack standards, making it hard to combine data sets.  Diagnostic companies are more likely than other healthcare manufacturers to be small companies with limited budgets.  The current cost of purchasing data, excluding analysis, is estimated to be between $25,000 and $200,000. CONCLUSIONS: Despite their potential, claims and EMR data sources have significant limitations in the detail they can provide related to diagnostic and lab services.  Additionally, big data is not affordabile for many diagnostic companies.  As a result, diagnostic companies face challenges in demonstrating both shortcomings of existing approaches and the clinical and cost utility of novel tests.  As diagnostics become more central to health decision-making and personalized medicine, data sources need to address existing limitations to better demonstrate their clinical and economic impact.