Introduction to Discrepancy and Reconciliation

Purpose

This document explains why different ad platform systems often report different counts and how those differences are operationally reconciled.

Key Takeaways

• discrepancy refers to differences between platform-side counts.
• reconciliation refers to the process of analyzing those differences and aligning an operational reporting basis.
• SSPs, DSPs, SDKs, and measurement vendors naturally observe events at different layers and timing points.

Basic Flow

Draft Structure

1. Why discrepancies happen

• timing differences
• render success criteria
• duplicate collection and timeout behavior

2. Why reconciliation matters

• to align billing and reporting
• to create a shared operational source of truth

3. Practical checkpoints

• the different roles of server logs and client logs
• where measurement vendor data enters the flow
• deduplication and idempotency rules

Prerequisite Concepts

• Understanding TrackingEvents, impression, click, and quartile
• Event Log Schema Basics

Next Document

• Understanding sellers.json and schain