Healthcare Revenue Cycle Analytics: $1.8M Underpayments Recovered Through AR Forecasting and Payer Benchmarking in Power BI
Numlytics delivered a complete healthcare revenue cycle analytics platform for a PE-backed RCM services company - building AR forecasting Power BI dashboards that achieve 93% accuracy at the 30-day horizon, a payer benchmarking intelligence layer that recovered $1.8M in systematic underpayments in year one, and a fully automated investor reporting engine that cut a 40-hour monthly manual process to under 2 hours.
The Challenge: A Healthcare Revenue Cycle Analytics Gap Costing Millions
A PE-backed RCM services company managing billing for multi-specialty physician groups across four East Coast sites had no healthcare revenue cycle analytics capability. Payer contract renegotiations were conducted without data. Provider clients were losing enterprise pitches to competitors who offered live AR forecasting Power BI portals. And systematic payer underpayments - from BCBS Illinois and Cigna HMO on Modifier 25 and outpatient surgery CPT codes had gone undetected for years.
- No payer benchmarking analytics: Contract renegotiations were conducted without CPT-level payment data. There was no system to compare actual reimbursements against contracted rates, meaning systematic underpayments from BCBS and Cigna went undetected month after month.
- No AR forecasting for provider clients: Provider clients had zero visibility into projected 30, 60, and 90-day collections. The company was losing enterprise pitches to competitors who could demonstrate live healthcare AR forecasting dashboards in Power BI during sales presentations.
- 40+ hours of manual investor reporting: The monthly investor pack was assembled manually across 4 sites and 3 billing systems - consuming senior staff time that should have been directed at collections, denial management, and client service.
- Fragmented claims data across three systems: Claims, ERA remittance, and eligibility data were siloed across AdvancedMD, Athenahealth, and manual Excel trackers making any cross-site or cross-payer analysis a multi-hour wrangling exercise with no single source of truth.
Without a functioning healthcare revenue cycle analytics layer, the company had no mechanism to detect underpayment patterns before dispute windows closed. The combined underpayment from BCBS Illinois and Cigna HMO both systematically paying below contracted rates on Modifier 25 and outpatient surgery codes had accumulated to $1.8M before Numlytics surfaced it through payer benchmarking analytics.
The Numlytics Healthcare Revenue Cycle Analytics Solution
Numlytics designed and built two integrated modules on Microsoft Fabric: a healthcare AR forecasting Power BI engine and a payer benchmarking intelligence layer, both feeding from a single unified data pipeline across AdvancedMD and Athenahealth, without disrupting any operational billing workflows.
-
01Unified Data Pipeline: AdvancedMD and Athenahealth Integration
Numlytics built Azure Data Factory pipelines ingesting claims data, ERA/835 remittance files, and eligibility records from both billing systems into a centralised Microsoft Fabric OneLake warehouse. Automated daily reconciliation catches discrepancies before they become billing patterns. This unified claims foundation is the prerequisite for any accurate healthcare revenue cycle analytics without it, AR forecasting models learn from the wrong data and payer benchmarking produces incorrect comparisons.
-
02Payer Benchmarking Analytics: CPT-Level Underpayment Detection
Numlytics built a Power BI payer benchmarking layer that compares each payer's actual payments against contracted rates at the individual CPT code level, surfacing systematic underpayments by payer, code, provider, and site. BCBS Illinois and Cigna HMO were flagged within the first reporting cycle for Modifier 25 and outpatient surgery underpayments. The payer benchmarking module provides the commercial team with precise, data-backed evidence before every contract negotiation - not instinct.
-
0330/60/90-Day Healthcare AR Forecasting Power BI Model
Numlytics built the AR forecasting model using historical payer payment velocity, claim aging distribution, and seasonal collection patterns from the Fabric Lakehouse. The model achieves 93% accuracy at the 30-day horizon - accurate enough for banking covenant compliance and board-level cash flow presentations. Provider clients now see live 30, 60, and 90-day collection forecasts in branded Power BI portals updated daily, a capability that directly won enterprise contracts previously lost to competitors.
-
04Investor and Board Reporting Automation via Power Automate
The 40+ hour monthly investor reporting pack previously assembled manually across 4 sites and 3 billing systems is now generated automatically via Power Automate from Microsoft Fabric on a scheduled trigger. Leadership reviews and approves it in under 2 hours. Power BI dashboard data populates investor decks, QBR presentations, and banking covenant reports with no manual intervention, eliminating the single biggest drain on senior staff capacity.
-
05Payer Contract Intelligence for Renegotiations
Numlytics built a contract intelligence module tracking each payer's payment performance trend over time enabling the commercial team to enter every renegotiation with granular evidence of underpayment frequency, affected CPT codes, and cumulative revenue impact. In the first renewal cycle post-launch, the team used the Numlytics payer benchmarking data to negotiate an improved BCBS Illinois contract term worth an estimated $400K in additional annual collections.
What Healthcare Revenue Cycle Analytics Delivered
- No payer underpayment detection - $1.8M lost undetected
- No AR forecasting losing enterprise pitches
- 40+ hour manual monthly investor reporting
- Claims data siloed across 3 disconnected systems
- Average AR days: 48 - no payer-level visibility
- $1.8M in underpayments recovered in year one
- 93% accurate 30/60/90-day AR forecasts in Power BI
- Investor reporting cut from 40 hours to under 2 hours
- Unified Fabric OneLake across all sites and systems
- Average AR days: 33 — payer-targeted collections strategy
Key Insights from This Healthcare Revenue Cycle Analytics Engagement
Every healthcare RCM company has the same surface-level data problem. What separates recoverable underpayments from permanent write-offs, and competitive AR forecasting from a spreadsheet, is a small number of structural decisions made before any dashboard is built.
Underpayment Detection Requires CPT-Level Granularity - Payer-Level Summaries Miss It
BCBS Illinois and Cigna HMO appeared to be performing acceptably at the aggregate level. The underpayment pattern only surfaced when payer benchmarking was applied at the CPT code level. Modifier 25 and outpatient surgery codes were systematically underpaid invisible in summary dashboards, clear in granular analytics. Any healthcare revenue cycle analytics layer that stops at payer totals will miss the underpayments most worth recovering.
AR Forecasting Accuracy Depends on Payer-Segmented Payment Velocity, Not Average Days
The 93% forecast accuracy came from building separate payment velocity models per payer, not from a single average AR days figure applied universally. BCBS pays differently than Cigna, which pays differently than Medicare. A single average-based AR forecast is accurate for no payer specifically. Payer-segmented modelling is what makes healthcare AR forecasting useful for banking covenants and board cash flow presentations rather than indicative but not actionable.
The AR Forecasting Capability Won More Enterprise Contracts Than Any Pricing Change
The most commercially impactful outcome of this healthcare revenue cycle analytics engagement was not internal, it was the 4× growth in average new enterprise contract value. Provider CFOs were not switching RCM partners based on pricing. They were switching because competitors could show them a live 30-day cash flow forecast in a branded Power BI portal during the sales meeting. The analytics capability was the product differentiator.
Unifying Billing Data Before Building Dashboards Is Non-Negotiable
The team had tried to build payer performance reports before Numlytics, but the AdvancedMD and Athenahealth data couldn't be reliably reconciled without a dedicated data engineering layer. Every analysis was a manual exercise. The Fabric OneLake pipeline built in step one of this engagement was not a preliminary task, it was the foundation that made every other healthcare revenue cycle analytics module accurate and maintainable. There is no shortcut past the data engineering work.