We investigate the optimal tea purchasing and blending problem at a global FMCG company. The company purchases tea leaves from various markets around the world and processes them in its global network of blending facilities. Tea leaves have different attributes such as taste, color and scent that vary depending on the producer and the season, and can be purchased at different price levels subject to market conditions. Leaves are mixed in blending facilities in appropriate quantities to achieve desired attributes of tea blend products. We formulate the optimal tea purchasing and blending problem as a multi-objective optimization problem. We also describe our experience in building a cloud-based decision support system based on our optimization model, which is used by the company on a weekly basis.