Overview
BSES — one of Delhi's most prominent electricity distribution companies, serving millions of households across Rajdhani and Yamuna zones — operates a dual-module SAP environment where financial data and operational meter data have historically lived in two separate worlds.
SAP R3, the core financial module, holds the accounting and asset view of every electric meter. SAP ISU, the Electricity Meter module, holds the operational and billing view of the same meters. With 14 million line items spread across both systems and years of independent operation, the two datasets had drifted apart — creating regulatory risk, asset register uncertainty, and a reconciliation challenge of extraordinary scale that had never been attempted before.
XLURSELF was engaged to do what no one had done: align both SAP modules completely, explain every discrepancy, and deliver a diagnostic report that could serve as a permanent reference for the finance and operations teams.
Challenge
The scale alone made this one of the most demanding reconciliation engagements we have undertaken. Fourteen million line items spanning electric meter records across two SAP modules — SAP R3 tracking meters as financial assets, SAP ISU tracking the same meters as live operational units serving customers. Both systems had evolved independently, accumulating discrepancies through data entry inconsistencies, meter lifecycle events recorded in one module but not the other, format mismatches across financial years, and operational changes that never fully propagated to the financial register.
The structural problem was compounded by the data itself: records exported in different formats across multiple financial years, heterogeneous field structures between R3 and ISU, and no existing process or tooling to bring the two datasets into a common frame for comparison. Before a single reconciliation query could run, the entire data foundation had to be rebuilt from scratch.
Solution
XLURSELF engineered a purpose-built reconciliation database that served as the neutral ground between SAP R3 and SAP ISU. All exports — spanning multiple financial years, multiple CSV formats, and both modules — were normalised into a unified schema, with ETL scripts validating referential integrity at every load step.
With a clean data foundation established, complex multi-layer SQL reconciliation queries matched records across both modules by meter serial number, location, financial year, and meter lifecycle status. Composite indexes were engineered specifically for the reconciliation join patterns — ensuring query performance remained consistent across the full 14-million-record dataset.
The result: a 98.59% reconciliation rate achieved in a remarkably short timeframe — and, critically, a comprehensive diagnostic analysis of every unreconciled record. The root causes of the remaining 1.41% variance were investigated, categorised, and documented in a formal investigative report — providing not just a reconciliation outcome but a permanent roadmap for resolving the underlying issues and preventing their recurrence.
How We Approached It
Data Inventory, Export Strategy & Format Analysis
Catalogued the full scope of SAP R3 and ISU data exports across all financial years — mapping field-level structure, format variations, meter lifecycle event types, and records existing in only one module. Defined the extraction strategy for both systems before a single line of data was processed.
Unified Reconciliation Database Design
Designed and built a SQL reconciliation database with a normalised schema capable of hosting both SAP modules in a single, comparable structure. ETL scripts transformed each year's export format into the unified schema with field-level validation and load-error logging at every step.
Composite Index Architecture for 14M-Record Performance
Engineered composite indexes on meter serial number, financial year, location code, and meter lifecycle status — the primary join keys across both modules. Index strategy ensured reconciliation queries executed efficiently across the full 14-million-line-item dataset without performance degradation.
Multi-Layer SQL Reconciliation Query Execution
Built and executed the core reconciliation SQL in passes: exact serial number matching first, then format-normalised matching for field inconsistencies, then proximity matching for partially captured records. Each pass produced a documented match confidence tier and a residual set for diagnostic analysis.
Diagnostic Analysis & Investigative Report
Performed a comprehensive root-cause analysis of every unreconciled record in the 1.41% variance set — categorising discrepancies by type: data entry errors, lifecycle events not propagated across modules, decommissioned meters not written off, and format-driven mismatches. Delivered a full investigative report with remediation recommendations, serving as a permanent reference for both finance and operations teams.
Impact
14 million
Line items reconciled
The largest data reconciliation XLURSELF has executed — 14 million electric meter records across two SAP modules harmonised for the first time in BSES's history.
98.59%
Data alignment achieved
An outstanding reconciliation rate for a dataset of this scale and complexity — achieved in a remarkably short timeframe through precise data engineering and SQL methodology.
100%
Variances root-caused
Every record in the unreconciled 1.41% was individually investigated, categorised, and documented. Not a single unexplained variance was left in the final report.
Roadmap
For future prevention
The diagnostic investigative report serves as a permanent reference — not just resolving today's discrepancies but illuminating the process gaps that allowed them to accumulate.
Key Outcomes
- First-ever reconciliation of SAP R3 (Financial) and SAP ISU (Meter) modules across BSES Rajdhani and BSES Yamuna
- 14 million line items processed through a purpose-built SQL reconciliation database with composite index architecture
- 98.59% data alignment achieved — the highest rate attainable given legitimate system lifecycle differences
- Comprehensive diagnostic report: every variance in the 1.41% gap categorised by root cause with remediation guidance
- Reusable reconciliation database and query framework enabling periodic recon runs going forward
- Investigative report adopted as the reference document for preempting future R3 ↔ ISU discrepancies
Facing a similar challenge?
Let's talk about what's broken in your finance operations. The first conversation is free, focused, and no-pitch.