Supercharging Dynamic Pricing for a Trucking Logistics Company

The client, a trucking logistics company, struggled to secure business buy-in to replace a legacy pricing model with one developed by its data science team. Traditional ML evaluation metrics don’t always reflect real-world performance, creating a bottleneck to deployment approval. The client turned to Aimpoint Digital to perform an extensive evaluation and certify the model’s code and architecture.

The Aimpoint Digital team worked closely with the client’s data science team to evaluate the model results, code, and architecture. We conducted architecture and code reviews through collaborative working sessions and independent analysis. Each step of the model pipeline and rules engine was reviewed and evaluated to verify the output was correct.

A detailed subset analysis was done on the model results to identify potential areas where the model performs poorly or might be overfitting. This analysis covered groupings of different locations, routes, distances, product types, and even customers. Visualizations were created to help identify patterns and communicate results to the business. We presented findings and recommendations to the founder, enabling an informed decision on live deployment.