Your client doubts the simulation model outcomes. How can you convince them of its validity?

To build trust: 1) The simulation should replicate with acceptable accuracy your customer's current outcomes under the current context. 2) The sensitivity analysis w.r.t. simulation variables and scenarios should produce "explainable" and "understandable" results, with as much drill-down support analysis as possible. This is especially important when the simulation results are not intuitive. A simulation has the advantage of transparency, as opposed to a black box. You have control over all the variables, so it is up to you to wisely display the right level of information to the user in order to achieve the right "explainability" of the outcomes.

Break down the model into understandable terms, explaining the key assumptions, inputs, and processes. Use visuals like flowcharts or diagrams to illustrate how the model works. Be upfront about the assumptions made in the model and why they are reasonable. Address any potential concerns by showing how the assumptions align with the real-world situation.

Primero es muy importante tener una comprensión exhaustiva del modelo utilizado, y también saber bien cuál es el perfil de la persona a la que queremos convencer del correcto funcionamiento del modelo. Esto nos dará herramientas que facilitarán el hallar las palabras y la cohesión adecuada para explicar de la manera más conveniente posible. Así, se nos hará más fácil explicar a la contraparte por qué la simulación realizada es razonable y adecuada para la representación que buscamos del fenómeno analizado.

You can use real data to validate the model and demonstrate its performance to your client or show that the results are close to those in the existing literature. You can also conduct an error analysis to prove that your simulations are efficient. It depends on the model.

When stakeholders question your model, it’s crucial to demonstrate its validity clearly. Begin with a transparent, step-by-step walkthrough of your model’s development, explaining why you chose this specific approach. Simplify complex processes into understandable terms and highlight your validation methods, including sensitivity analysis to illustrate how different inputs impact outputs. Conclude with a live demonstration to showcase the model’s practical application and effectiveness. This thorough explanation will help alleviate concerns by showcasing the model’s reliability and your meticulous engineering approach.

Realizar un backtesting del modelo con datos históricos y verificar que las predicciones del modelo son consistentes con lo ocurrido durante ese período es una buena manera de validarlo

Una de las formas de converser a un cliente sobre la validez de un modelo de simulación en la etapa de pruebas de escenarios es : 1. Simular escenarios y condiciones actuales buscando demostrar lo acertado que está la simulación sobre la respuesta obtenida. 2. Mostrar los resultados de la simulación para casos poco comunes evidenciando el porcentage de fiabilidad/ajuste entre la respuesta de la simulación vs lo medio/obtenido en la realidad. Si generamos un marco de referencia para el cliente con casos anteriores y cuantificamos la fiabilidad del modelo en todos estos casos, podremos despejar cualquier duda que tenga de la simulación en escenarios diversos.

All the responses here seem to assume that the model being presented is correct. If your client doubts your model, it's usually for good reason. More time needs to be spent with the client understanding the reality of the problem being represented.