A metric engine is to business measurement what a search engine is to text retrieval — but instead of finding documents, it measures positions. Unlike a reporting system that computes balances on demand from raw transaction data, a metric engine pre-summarizes against all meaningful combinations of keys at posting time, making any report instantaneous regardless of data volume.
Written from 30 years of financial systems consulting experience, this monograph explains why current financial reporting architectures carry far more cost than necessary — and what a purpose-built metric engine architecture looks like in practice.
📄 The full PDF edition and accompanying slide decks are available to Sharealedger members (free membership).
Contents
Preface & Introduction
- Quantification: The Basis of Business Measurement
- The Assembly of Data: A Manufactured Goods Warehouse Analogy
The Historical Process
- Obstructions: Impediments to Accuracy, Completeness, Transparency, Timeliness, Availability, and Cost-Effectiveness
- Origins of Understanding: Transactions to Balances
- Clarity: Transparency, Traceability and Repeatability
- Posting and Reconciliation: The Original Business Systems
- Proliferation: Duplicative Data Supply Chains
- Time Zones and Clock Speeds: The Periodicity of Reporting
An Alternative Approach
- JIT Manufacturing: Just-In-Time Analysis
- The Metric Engine: Trusted, Aggregated, Structured Data
- Gathering Transactions: Data Supply Chain Part 1
- Low-Level Posting: Data Supply Chain Part 2
- High-Level Aggregation: Data Supply Chain Part 3
- Scale: Producing the Goods
Conclusions
- Data Reactive Functions: Major Data Supply Chains
- Consolidation: The Impact of Change
Epilogue
Key Concepts
Step-Ups
Pre-summarizing transaction data at multiple levels of a hierarchy simultaneously — so any balance at any level is already computed, never derived on demand.
Temporal Events
Handling time as a first-class data attribute — enabling point-in-time reporting, period-end snapshots, and retrospective analysis without reprocessing.
The Periodicity Problem
Why batch overnight processing exists, what it costs, and how a metric engine architecture eliminates the need for nightly batch posting runs.
GenevaERS in Practice
IBM's Scalable Architecture for Financial Reporting (SAFR / GenevaERS) as a real-world implementation of metric engine principles at enterprise scale.
Related Episodes
The Metric Engine series walks through these concepts episode by episode.