Cancellation Prediction for a Real Estate–Financial Company

The Challenge

The company faced high cancellation levels in the early stages of the customer lifecycle, with an average impact close to 45% of the contracted amount.

The analysis relied only on traditional financial variables, with no real ability to anticipate cancellations.

The Solution

Infomedia implemented a predictive model that integrated:

• Machine learning
• Financial and operational variables
• A geostatistical component to infer socioeconomic level
• Actionable risk segmentation

In just 12 weeks, the organization moved from uniform strategies to targeted interventions by risk level.

• Customers with a probability ≤ 0.4 represented only 12% of the risk
• Customer prioritization by economic impact
• Greater efficiency in resource allocation
• The operation evolved from reactive to preventive.