Change Data Capture (CDC) is a set of software design patterns used to determine and track the data that has changed in a database so that actions can be taken using those changes. For CRM teams, CDC is the difference between a system that reflects reality and one that reflects yesterday.
How CDC Works
Traditional data integration approaches rely on periodic queries: every hour, every night, or every few minutes, a job runs and checks for changes. This creates inherent latency. CDC takes a fundamentally different approach — instead of polling the database, it reads the database transaction log directly.
Every relational database maintains a write-ahead log (WAL) or binary log that records every single change made to the database in strict order. CDC tools tap into this log stream and capture inserts, updates, and deletes the instant they are written — before any delay is introduced.
The CRM Sync Problem
CRMs are only as useful as the data they contain. When a customer upgrades their plan, does your CRM know within seconds or overnight? When a deal closes in your billing system, does the opportunity status in Salesforce update immediately or after the next batch job?
The gap between what happened and what your CRM knows is called data lag. For sales teams, data lag means calling on outdated context. For marketing, it means firing the wrong campaign. For customer success, it means missing a churn signal until it is too late to act.
Why Polling Approaches Fail at Scale
The traditional approach to CRM sync — querying source tables for changes since the last run — has three fundamental problems:
- Latency scales with query interval. You cannot safely poll every second without overloading production databases.
- Deletes are invisible. A deleted row leaves no trace for a polling query to detect.
- High watermarks miss non-monotonic updates. If a timestamp column is not updated consistently, rows are skipped.
CDC Changes the Equation
Log-based CDC eliminates all three problems. Because it reads the WAL directly, it captures every change including deletes, in strict order, with zero additional load on the production database. Events arrive within milliseconds of the original write.
For CRM sync specifically, this means your CRM record is updated the moment the source database record changes. No polling interval. No missed deletes. No gaps.
Getting Started with CDC for CRM
Implementing CDC requires a connector that can read your database log and a delivery mechanism that can write to your CRM. Salmon Labs provides both as a managed service — connect your database, configure your CRM mapping, and your sync is live in under two hours.
The result is a CRM that operates on data that is always current, always accurate, and always actionable.