Traditional behavioral scoring has filled the bill for many years by providing a basis for making decisions about which delinquent accounts to call. But with millions of dollars in uncollected delinquent dollars, many financial institutions are supplementing their behavioral scoring approach with action-specific modeling solutions that predict the response to particular actions – such as sending a letter, making a call, and placement with an agency – on specific accounts. The most appropriate and profitable action may then be taken for each account.

The benefits of action-specific modeling include: an improved cure rate, resulting in more dollars collected along with lower roll rates and fewer chargeoffs; reduced call volume, making better use of collection call center resources; improved customer retention, by eliminating collection calls to those likely to pay without a call; and fast return on investment, with payback typically in just a few months.

In the early stages of delinquency, the percent of self-cures (those likely to pay even without a call) can be 30% to 50% or higher. By accurately identifying self-cures, action-specific modeling helps collection centers focus activity. For late stage delinquency, it can help identify accounts that will not pay unless a collection agency calls. An early decision to place these accounts externally could mean collecting while the debtor still has funds available.

Behavioral scoring uses a statistical method to predict the likelihood that an account will make a payment, reach a later stage of delinquency, or eventually chargeoff. The traditional strategy classifies delinquent accounts into high-risk, medium-risk, and low-risk categories based on the value of a behavioral score. Collection call centers focus their efforts on high-risk accounts. As a result, high-risk accounts are called, while low-risk accounts are not.

However, behavioral scoring is limited. First, it does not identify which accounts will respond to a call and which will not. Models are developed on an average set of actions without considering which actions are effective on a particular account. Second, traditional behavioral scoring uses a group-based approach. All accounts in a high-risk group are assigned a specific action, such as a phone call, while all accounts in a low-risk group are assigned another action, such as no call. The collection call center misses the opportunity to call low-risk accounts that need agent assistance to pay, while needlessly calling those who will pay without a call or those who will not pay no matter what actions are taken. These inefficiencies limit the amount that can be collected, and waste call center resources.

Action-specific modeling, on the other hand, allows collection call centers to see the whole picture by predicting the response to collection actions on each account. Both self-cures and non-cures are identified, so resources can be directed to responsive accounts.

Action-specific models are built using the creditor’s data, such as:

    • Account master file: delinquency history, geography, time on file, balance
    • Relationship: breadth and depth of relationship
    • Collection activity: payment history, calls, promises, broken promises.

Models that predict the outcomes of certain actions are developed using an experiment to eliminate the bias in data. Each action (call vs. don’t call; low intensity calls vs. high intensity calls; settlement offer vs. agency; active skip trace effort vs. letter) is applied to a random sample of accounts, and the responses are tracked over the performance period. Since the samples are random, the models can be applied to accounts across the population.

Action-specific modeling software continually captures new and updated data. Periodic experiments and model rebuilding ensure that changes in the portfolio are captured and new predictive characteristics are included.

Through this ongoing process, collection call centers can continually test new actions and new populations to refine and enhance their strategies and resource allocations.

“Lois Brown is vice president of marketing at Austin Logistics, which develops and markets analytic software applications that maximize call center effectiveness, consumer collection, and risk management.”

About Kollect Systems

Kollect Systems is an innovative tech platform provider with BankTech and FinTech software solutions which leverage AI based decisioning and workflow technologies to help lenders perform Debt Collections & Recovery (BankTech) processes effectively and for mid-size to large scale enterprise companies (FinTech), to automate Receivables, e-Invoicing & Payments better.

Kollect’s Solutions :

    • KollectApps for Lenders (BankTech) 
    • KollectValley for Finance (FinTech)
    • KollectRepo 
    • Data Integration & Analytics

Schedule a Free Consultation

Got-a-question-Yellow

Got a Question?

We’d love to hear from you. Send us a message and we’ll respond as soon as possible.

Kollect Systems is an innovative tech platform provider with BankTech and FinTech software solutions which leverage AI based decisioning and workflow technologies to help lenders perform Debt Collections & Recovery (BankTech) processes effectively and for mid-size to large scale enterprise companies (FinTech), to automate Receivables, e-Invoicing & Payments better.

Office

Copyright © 2022 Kollect Systems Sdn. Bhd.      |      Privacy Policy