Capital One rolled out check imaging by modeling who each plan would hurt

Capital One planned to deploy check imaging technology to tellers across roughly 1,000 retail branches. The obvious project-management answer was to pick the fastest or cheapest rollout path, but similar bank projects had suffered delays and dissatisfied customers and employees.

The first decision model gave management an overall quantitative comparison of rollout alternatives. Executives then noticed the trap: an option that looked fine in aggregate could still damage one business area badly enough to stall the project.

The team rebuilt the analysis around group decision theory, separating the view of each stakeholder group from the combined company view. That exposed why all three initial alternatives were flawed and helped executives design a fourth rollout strategy.

The elegant move was making organizational friction an input to the model. Instead of treating resistance as a post-model change-management chore, Capital One used it to shape the implementation plan itself.

Optimize adoption pain, not just rollout speed

Averaging stakeholder utility would have hidden local pain points that could delay adoption.

The method turned a rollout argument into a structured comparison of trade-offs by business area.

The final strategy emerged from the model's failures, not from forcing one preset option to win.

For big internal rollouts, the clever model is often the one that shows whose workflow breaks, not the one that maximizes an aggregate score.

The paper reports that the new strategy avoided delays seen at other banks implementing similar technology, saved approximately $30 million, and improved customer and employee satisfaction. The case was published in INFORMS Journal on Applied Analytics in 2023.