Incremental Recovery Rate (IRR)
1) Meaning
Incremental Recovery Rate (IRR) is a causal effect metric that measures how much more recovery (or repayment, or conversion back to “good status”) happens in a treatment group compared to a control group.
In other words:
“What fraction of people recovered because of the intervention, compared to if we had done nothing?”
This metric is widely used in collections/recovery, healthcare, and customer churn management.
2) Formula
$\text{Incremental Recovery Rate} = P(\text{Recovery} \mid \text{Treatment}) – P(\text{Recovery} \mid \text{Control})$
Where:
- $P(\text{Recovery} \mid \text{Treatment})$ = recovery proportion in treatment group
- $P(\text{Recovery} \mid \text{Control})$ = recovery proportion in control group
3) Example
Suppose a bank tests a new collections strategy:
- Treatment group (contacted with new script): 35% repayment
- Control group (no new script): 28% repayment
$\text{Incremental Recovery Rate} = 35\% – 28\% = 7\%$
The new strategy increased recovery by 7 percentage points compared to the old process.
4) Relation to Other Metrics
- Incremental Sales: Same idea, but applied to sales instead of recoveries.
- Uplift / Incremental Effect: Incremental recovery rate is essentially an uplift metric for recovery outcomes.
- Total Incremental Recoveries: If population = 10,000 delinquent accounts, 7% extra recovery = 700 accounts recovered.
5) Why It Matters
- True effectiveness measure: Shows the net effect of an intervention, not just the overall recovery rate.
- ROI focus: Distinguishes between organic recoveries (would have happened anyway) vs. those driven by the strategy.
- Targeting optimization: With uplift models, you can find which customers are most likely to show incremental recovery.
- Policy evaluation: Used in healthcare (patients recovering due to treatment vs. placebo).
Bottom line:
Incremental Recovery Rate = the additional recovery percentage directly caused by an intervention (treatment vs. control). It’s the “uplift” metric applied in collections, healthcare, or retention contexts.
Discover more from Insightful Data Lab
Subscribe to get the latest posts sent to your email.
