Reverse ETL Overview

Reverse ETL integrations in Matia move data from your data warehouse (or other source) into operational tools — CRMs, ad platforms, support tools, and other SaaS destinations. Each integration uses a model (SQL that defines the dataset) and a destination object (e.g. Contact, Company) with field mapping and a sync mode.

What is a Reverse ETL Integration?

An integration is one reverse ETL pipeline: a configured connection from a source (typically your warehouse) to a destination (e.g. HubSpot, Salesforce). You define what to sync with a model (SQL query) and where it lands with a destination object and field mapping. Each sync pushes changed records (inserts, updates, deletes) according to the sync mode. Matia tracks successful, rejected, and invalid records per run.

How Data Flows

1. Model: A SQL query selects rows and columns from your warehouse. The result set is the data set for the sync.
2. Sync: Matia compares the model result to the destination (using primary keys and sync mode) to determine changed records — inserts, updates, deletes.
3. Push: Each changed record is sent to the destination API. Successful records are written; rejected records are refused by the destination (API request/response visible for debugging); invalid records are flagged by Matia (e.g. duplicate primary key) and not sent.

Managing Reverse ETL Integrations

• Integrations (sidebar): List of all integrations; filter by reverse ETL. Add Integration and choose a warehouse source and operational destination.
• Integration details: Status (sync history, changed records, successful/rejected/invalid breakdown), Settings (model SQL, destination object, sync mode, field mapping, trigger, notifications, test, full reset, delete), Changelog (audit log). Reverse ETL integrations do not have a Schema or Schema changes tab.

Key Concepts

• Model: The SQL that defines the synced dataset. Preview and edit in Settings.
• Destination object: The object in the destination that receives rows (e.g. Contact, Company, custom object).
• Sync mode: Mirror (inserts + updates + deletes), Upsert (inserts + updates), or Update only / Delete only where supported. See Sync modes and destinations.
• Field mapping: Primary keys (source and destination) and column-to-field mappings; auto-mapping can suggest mappings from column names.

For a first pipeline, see Create your first reverse ETL integration. For reference, see How reverse ETL works, Sync modes and destinations, and Models and destination objects.