How to set up a data pipeline with Fivetran

How to Set Up a Data Pipeline with Fivetran

Fivetran is a managed ELT (Extract, Load, Transform) platform that replicates data from sources to cloud data warehouses without custom code. As of April 2026, Fivetran supports 500+ connectors and integrates with Snowflake, BigQuery, Redshift, Databricks, and other destinations. Pricing is usage-based on Monthly Active Rows (MAR), starting at $1/credit on the Standard plan.

Step 1: Add a Connector

Log into the Fivetran dashboard and click "Add Connector." Search for the data source from the connector library:

• SaaS applications: Salesforce, HubSpot, Stripe, Shopify, Google Analytics, Jira
• Databases: PostgreSQL, MySQL, SQL Server, MongoDB, Oracle
• Files: S3, Google Cloud Storage, Azure Blob, SFTP
• Events: Segment, Webhooks, Kafka

Step 2: Authenticate the Source

Each connector has a specific authentication method:

• OAuth (Salesforce, HubSpot, Google): Click "Authorize" and sign in
• API key (Stripe, Intercom): Paste the API key from the source application
• Database credentials (PostgreSQL, MySQL): Provide host, port, database name, username, and password. Fivetran requires read-only access and recommends a dedicated service account

Fivetran provides setup guides for each connector with exact permissions and configuration steps.

Step 3: Select a Destination

Choose the data warehouse where Fivetran will load the data. Configure the destination once (Fivetran remembers it for all connectors):

• Snowflake: Provide account URL, database, warehouse, and role
• BigQuery: Provide project ID and dataset (Fivetran creates tables automatically)
• Redshift: Provide cluster endpoint, database, and credentials
• Databricks: Provide workspace URL, token, and catalog

Step 4: Choose Schema and Tables

Fivetran displays the source schema and allows selection of which tables to sync. Options:

• Sync all tables — Replicate the entire source schema
• Select specific tables — Choose only the tables needed
• Column hashing — Hash sensitive columns (PII) during replication

Fivetran maintains schema consistency: if the source adds a column, Fivetran automatically adds it to the destination.

Step 5: Set Sync Frequency

Configure how often Fivetran checks for new and changed data:

• 5 minutes — Near real-time (highest MAR usage)
• 15 minutes — Balanced freshness and cost
• 1 hour — Standard for reporting workloads
• 6 hours / 24 hours — Batch analytics with lower cost

Step 6: Start the Initial Sync

Click "Save & Test" to validate the connection, then start the initial sync. The first sync replicates all historical data and may take hours or days depending on data volume. Subsequent syncs are incremental, transferring only new and changed records.

Post-Setup: Connect to dbt for Transformations

Fivetran loads raw data into the warehouse. Use dbt (data build tool) to transform raw tables into analytics-ready models. Fivetran offers a native dbt Cloud integration that triggers dbt runs automatically after each sync completes.