Big-data pipeline
Clearing files and real-time authorizations processed at billions-of-rows scale.
Payments · Canada
Dormant-card reactivation
Behavioral targeting over a multi-billion transaction history for a global payments provider — the right offer, attached to the card, at the moment of purchase.
24% of dormant cards re-activated in 60 days — vs 6% control
2B+ transactions processed for behavioral clustering
$4.1M annual interchange lift for the issuer
<48 h campaign creation — down from 3 weeks
The challenge
What was in the way.
- Across two billion historic card transactions, 12–15% of active cards had gone dormant — no purchase in 90 days.
- Dormancy erodes interchange revenue and strands reward points at global scale.
What we built
The system, in brief.
Behavioral clustering
Every cardholder grouped by spend velocity
Offer recommender
A gradient-boosted model predicts the offer most likely to trigger reactivation
Card-linked delivery
When a dormant cardholder transacts anywhere
Test-vs-control analytics
Incremental lift
Outcomes
What changed.
- 24% of dormant cards re-activated within 60 days, against 6% in the control group.
- Reactivated cardholders generated $112 of average incremental spend in the first 90 days — about $4.1M in annual issuer lift.
- Coupon redemption multiplied against generic broadcast offers.
- Campaign creation fell from three weeks to under 48 hours via a self-serve dashboard; fully PCI-DSS compliant.
Client referenced by sector and country · detailed references on request