Carlos Armando Zetina: “The model is an approximation for a business’s decision problem because it inevitably needs to make simplifying assumptions such as limited scope, finite horizons, deterministic parameters, and cost approximations to be amenable to solving. Since the optimality gap is a measure based on the model, using it to measure a solution’s success is misleading.

What matters to the business is its Key Performance Indicators (KPIs) e.g. out of stock, actual transportation costs, lead times, revenue, profit margins and other KPIs the solution can directly impact. These KPIs should be tracked continuously before and after the optimization solution is in production, guiding model adjustments to make decisions that align better with the ultimate business goals. Optimization solutions in industry are living entities that require continuous improvement and KPI monitoring. Solving the model is only the beginning of the decision science process.” Link