This case study describes a retail marketing mix optimization. Key inputs are marketing mix models, marketing cost functions, and client business rules. Outputs are recommended prices and ad schedules. Models, cost functions, and business rules constitute a large-scale MINLP not easy to solve in practice, but we find a reasonable piecewise linear approximation that can be solved with MIQP, which is more manageable. The optimization delivers a 5%-10% improvement in revenue and profit plus a significant ad cost savings opportunity.