Case Studies · FinTech

Automated billing analytics for a debt-collections platform.

A leading debt-collections technology platform billed 300+ lenders across SMS, WhatsApp, IVR, email and legal notices — every invoice computed by hand from six fragmented databases. We built a Redshift billing data mart and automated the whole thing.

Industry
FinTech / Debt Collections
Technology
AWS · Redshift · EMR
Duration
12 weeks
Scale
300+ lender clients
The business problem

Billing that couldn't scale.

As the client grew, the finance team hit an unsustainable bottleneck. Each client ran a distinct pricing structure, and every month meant manually merging pricing with usage from disparate source systems — slow, error-prone, and impossible to scale.

Manual billing

Hours of manual cost computation every month across 300+ clients, each with unique pricing and usage patterns.

Complex pricing

Slab-based and count-based tiers with effective date ranges per client, plus monthly license fees — non-trivial to automate.

Fragmented data

Six-plus separate databases for SMS, WhatsApp, IVR, email and notices, with no unified view of cost or usage.

No self-service

Finance and business teams couldn't independently slice, drill down, or compare costs and usage across clients and channels.

Scalability gap

Every new channel or pricing tier meant reworking a manual process — growth made the problem worse, not better.

Attribution complexity

Multi-file loan allocations, DPD bucket movements and campaign attribution layered extra complexity onto every calculation.

The solution

A dedicated billing data mart, fed daily.

An end-to-end platform on the client's existing AWS estate, built on three pillars: a purpose-built Redshift data mart, automated ETL, and pre-computed, dashboard-ready metrics.

Billing data mart

Three denormalized Redshift tables — Usage, Pricing and Cost — capturing day-level metrics across every channel, per client, built for query speed and minimal joins.

RedshiftUsage · Pricing · Cost

Automated ETL

MWAA-orchestrated EMR Spark pipelines extract from six sources, apply each client's pricing logic, and compute cost metrics on a daily schedule — handling multi-file allocations and DPD transitions.

EMR SparkMWAAS3

Dashboard-ready metrics

Pre-computed KPIs across Cost, Usage and Gross Margin — structured around six dimensions and ready to plug into QuickSight or any BI tool to publish dashboards rapidly.

QuickSightCost · Usage · Margin

Source Systems

SMSDatabase
WhatsAppDatabase
IVR / DiallerDatabase
EmailDatabase
NoticesPhysical & digital
Allocation & PricingFiles & config

Ingestion & ETL Layer

Amazon S3 Data LakeRaw & staged data
AWS MWAAOrchestration
Amazon EMR (Spark)ETL & transforms

Amazon Redshift — Billing Data Mart

Usage TableChannel-wise daily usage
Pricing TableSlab & count rules + dates
Cost TableUsage × pricing = cost

Dashboard-Ready Metrics — plug into any BI tool

Cost MetricsChannel · client · time-series
Usage MetricsVolumes · module · trends
Gross MarginProfitability by client/channel

Dimensions: Time (daily / weekly / monthly) · Module · Client · Loan Type · Region · DPD Bucket

Business benefits

Hours of manual work, gone daily.

70%+
Reduction in manual effort
99.9%
Data accuracy vs. manual
Daily
Automated cost refresh
6
KPI dimensions, dashboard-ready
BeforeAfter
Manual monthly billing calculationsFully automated daily cost computation
Fragmented data across 6+ systemsUnified billing data mart in Redshift
No structured cost or usage metricsPre-computed, dashboard-ready KPIs
Weeks to onboard new pricing tiersExtensible schema for rapid additions
No gross margin visibilityMargin metrics ready for instant visualization
Technology stack
Amazon S3Amazon RedshiftAmazon EMR (Spark)AWS MWAA (Managed Airflow)Amazon QuickSight12-week deliveryAgile sprints
Your billing, automated

Still computing invoices by hand?

If complex, per-client pricing is eating your finance team's month, we can build the same kind of automated data mart for you. Let's scope it.

Book a consultation →