GM swapped gut-feel trim and price decisions for a consumer choice model

Contenting and pricing decisions define a car's trims, features and prices, and they are locked in years before launch, so a wrong call costs hundreds of millions in tooling and lost sales.

GM's Vehicle Content Optimization replaced that intuition with data: discrete choice models estimate how customers value each feature, and novel optimization algorithms search the space of possible order guides for the one that best balances customer value, price and profit.

GM Finance verified that VCO enabled $4.4 billion of incremental lifecycle profit and that it has been applied to more than 85 vehicle programs worldwide since it was integrated into the development process. The work was a 2022 Franz Edelman Award finalist.

Optimize what each trim includes, not what to build

Choice models reveal what customers will pay for instead of what planners assume they will.

Multi-objective search balances customer value, price and cost instead of maximizing one at the expense of others.

Deliberate contenting avoids over-tooling features nobody buys and under-offering ones customers want.

A systemized decision makes each new vehicle program repeatable instead of starting from scratch.

When a product has thousands of configurable points, modeling customer choice beats trusting a team's feel for what will sell.

VCO became part of GM's global vehicle development process and grew into a platform used for complexity reduction and electric-vehicle forecasting. GM applied the same choice-modeling machinery to longer-term EV portfolio and pricing questions as it pursued its zero-emissions goal.