Ledge Reconciliation Engine Overview

Introduction

This document serves as a starter's guide to the Ledge Reconciliation Engine, used for transaction-level reconciliation across any data source. The input of the reconciliation engine are transactions from different data sources, and the outputs is a dataset providing transaction-level matching information, GL updates (based on Ledgering Rules), and Payment Story visualizations.

Supported Data Sources

Ledge supports a comprehensive range of data sources, including internal systems and databases, bank statements, reports from PSPs, data from check and cash processors, information from spend management systems as well as arbitrary data sources (e.g. custom files or reports). For information on supported formats, data transfer procedures, and an overview of the ingestion process, please refer to the Ledge Data Ingestion Overview.

Workflows and Matching Rules

Workflows are business process codified as queries on the payment story graph. A transaction may participate in multiple workflows, determining its reconciliation status. Workflows are a composition of Data Sources and Matching Rules, that describes a business process.

Users have the capability to define and fine-tune Workflows and Matching Rules. This empowers accounting, finance operations, and bookkeeping teams to create Workflows that align with their established processes and set rules that ensure accurate matching and validation of financial data.

Ledge's matching logic is based on a collection of reconciliation patterns and heuristics, and is customizable for different business needs. Patterns are especially common in scenarios involving standard data sources, like a refund connected to the original payment on a PSP charge & refund scenario. Some of the key reconciliation heuristics can be based on exact or fuzzy matching, historical patterns, priority based matching, etc.

Payment Story

The Payment Story serves as a powerful tool for providing a transparent and comprehensive visualization of a transaction's lifecycle from initiation to completion.This visualization not only showcases the journey of an individual transaction but also highlights its connections to other related transactions.

Matching Types

The Ledge Reconciliation Engine offers support for a range of transaction-matching models, including:

One-to-One Matching: This model involves matching one item from the first dataset to one item in the second dataset. It's commonly used for straightforward reconciliations where each item has a single corresponding match.

One-to-Many Matching: In this model, one item from the first dataset can be matched to multiple items in the second dataset. This is useful in scenarios where one record can have multiple related matches.

Many-to-One Matching: Conversely, many items from the first dataset can be matched to a single item in the second dataset. This model is employed when aggregating or consolidating data from multiple sources into a single record.

Many-to-Many Matching: In the many-to-many model, multiple items from both datasets can be matched to multiple items in the other dataset. It's used in complex scenarios where there are multiple possible matches between the datasets.