Credit Collection


Credit Collection is designed to improve the profitability of organizations in economical downturn period. The main objective is to maximize the value of the credits recovered and grow the chance of return of credits on the individual customer. This Advanced Analytic Application allows the planning of the best contacts, optimizing the channel lists and operator, identifying the names to be managed by third parties.

Business Context

Despite attention by banks when granting credit facilities, defaulting customers are on the rise, which is why we are witnessing a stepping up of – direct and indirect – recovery activities, implemented to mitigate losses, which are starting to become huge. The recovery actions, strategies, and channels, however, all have different costs and success rates.  To recover the greatest share of receivables possible while maintaining a sustainable level of costs, it is important to direct customers to the best channel, secure in the knowledge that you have approached each customer in the best way possible. Predictive models play a fundamental role in this industry, on the one hand enabling users to correctly predict the likelihood of repayment and on the other, optimising the lists of contacts, operators, and channels.

ì4C Application

The Credit Collection application allows users to establish business objectives in terms of expected repayments and affordable costs, and to set up information about the costs of use of the various recovery channels.  The application uses analytics, with daily indicator updates, to allocate each customer a potential recovery score and the recovery channel identified as optimal. With these indicators, the user can create an overview of the status of its receivables at likely repayment and have constantly updated lists of recovery customers at their disposal at all times. This way, they can make the best possible use of the contacts on the list, in relation to the objectives and constraints set.

Business Pain – Key Features

  • Improve Credit Collection efficiency
  • Predict Customers most likely to repay
  • State of the art predictive methods
  • Logistic Regressions
  • CRT, QUEST, CHAID trees
  • Outliers detection
  • Channel/action optimization
  • Reduce the need for statistical skills
  • Automated data preparation
  • Map channels and actions
  • Automated score refreshing
  • Automated outlier detection
  • Wizard assisted modeling
  • Automated modeling
  • Freedom and flexibility for users with skills
  • Not a black box, access to models
  • Drivers creation and selection
  • Default parameters for less experienced analysts
  • Hold out for testing
  • Partitioning/sampling
  • Lift chart/gains chart
  • Managing the whole decision process, till performance measurement
  • Action definition
  • Model identification
  • Deployment
  • Performance measurement
  • ROI for credit collection
  • Credit collection success ratio insights
  • Automate process
  • Create predictive projects
  • Schedule model update
  • Schedule score update
  • List of contact for channel