Recovery Intelligence

From Propensity Scores to Recovery Decisions

Scores sitting in a column don't collect. DebtorIndex maps every tier to a treatment and puts it on tomorrow's call list.

Score, decide, act, measure. Skip a step and the score is just a number.

Summary

Scoring only pays when each score drives a treatment: call, text, skip trace, hold or wait.

Expected recovery value (likelihood times collectible amount minus cost) beats raw likelihood.

DebtorIndex turns tiers into work lists and feeds the daily call list builder, with logged overrides.

Key takeaways

Write down the treatment for every tier.

Build lists and audiences from tiers, not age.

See DebtorIndex in a demo.

The gap between model and money

Lots of agencies bought scoring. Few can say what changed. The score became a column and collectors kept working the old way.

Expected recovery value

Payment likelihood alone can mislead. A 70 percent chance of collecting $40 is worth less than a 30 percent chance of collecting $2,000. A more useful measure is expected recovery value: likelihood multiplied by the realistic collectible amount, minus the expected cost of working the account.

That calculation naturally answers questions like:

Is this account worth a collector's call, or is a low-cost text the right first touch?
Is a bureau pull justified here?
Should this account wait while higher-value work is done?

Map tiers to treatments

Write down what each tier means operationally. For example:

Tier A: call today; collector with the best fit; follow up within 48 hours.
Tier B: voicemail and text sequence, then a call if they respond.
Tier C: low-cost digital outreach; review monthly.
Tier D: minimal effort unless new signals appear.
Holds: never contacted, regardless of score.
Fix contact data: skip trace before any more attempts.

Then make sure call lists and campaign audiences are actually built from those tiers.

Measure tiers against reality

Every month, compare each tier's predicted performance with actual payments. If tier B is outperforming tier A, something is off: the model, the treatment, or the data. Recalibrate when your portfolio mix changes, for example after taking on a new client or debt type.

Keep humans in the loop

Supervisors will sometimes know something the model doesn't. Allow overrides, but require a reason and record who made them, so you can learn from them and keep decisions auditable.

Watch the economics

Track labor per resolved account, recovery by tier and unnecessary data spend. The point of prioritization is to get more recovered per hour and per dollar.

How DebtorIndex does it

DebtorIndex scores accounts from their real behavior (payments, contacts, promises, plans and responses across channels), ranks them A to D, suggests the best way and time to reach each person and builds Call, Fix Contact Data, Bureau Pull, Holds and Legal lists. Tiers feed the daily call list builder. Overrides need a reason and are logged with the user's name.

The bottom line

Turn forecasts into treatments and treatments into lists. That's what DebtorIndex is for, and it's available for demonstration now.

See it working in your agency

Book a 30-minute guided Nexum demo built around your agency's real bottleneck, or call 502ZONE directly at (786) 524-4252. Nexum Open CRM is free: no license fee, no seat fees, no subscription. Add 502ZONE modules only when you want more recovery.

Bring us your most expensive collections problem.

Our guided 30-minute session focuses on your real workflow and the problem costing you the most.

Schedule a Demo