How to Set Up Fivetran to Sync Data to Your Warehouse

Fivetran moves data from your tools into a central warehouse so your team can query everything from one place. Instead of writing and maintaining custom scripts for each source, you configure connectors through a web interface and Fivetran handles the rest: schema detection, incremental syncing, error recovery, and schema drift management.

According to Fivetran's own benchmarks, their connectors typically sync only new or changed records rather than reloading entire datasets each time, which cuts sync windows by 60 to 90 percent compared to full-load pipelines.

How Fivetran Works

Fivetran sits between your source systems and your data warehouse. It connects to a source on a schedule, pulls new or updated records using change data capture or API polling, normalizes the schema, and loads the result into your destination. Schema changes in the source, such as a new column in a Salesforce object, are detected automatically and propagated to your warehouse without manual intervention.

The architecture has three parts: connectors (configured per source), a destination (your warehouse), and a transformation layer (optional, powered by dbt). You do not need to manage infrastructure for any of these.

Step 1: Create Your Account and Choose a Plan

Go to fivetran.com and sign up. Fivetran offers a 14-day free trial that includes access to all connector types and up to five connectors. After the trial, pricing is based on monthly active rows, which are rows that were inserted or updated during the billing period.

For most small teams, the Starter plan covers 500,000 monthly active rows. A mid-size SaaS company syncing Salesforce, HubSpot, and Stripe typically falls in the 1 to 5 million MAR range per month.

Once your account is active, you land in the Fivetran dashboard. The first thing to configure is your destination.

Step 2: Configure Your Destination

Navigate to Destinations in the left sidebar and click Add destination. Fivetran supports Snowflake, BigQuery, Databricks, Redshift, Azure Synapse, and DuckDB, among others.

For Snowflake (the most common choice for mid-market companies):

1. Select Snowflake from the destination list
2. Enter your Snowflake account identifier, warehouse name, database, and schema
3. Create a dedicated Fivetran user in Snowflake with the permissions Fivetran specifies — the setup guide in the dashboard includes the exact SQL commands to run
4. Enter the credentials and click Save and Test

Fivetran will verify connectivity before you proceed. If the test fails, the error message usually specifies the missing permission or incorrect credential field.

For BigQuery, the setup uses a service account JSON key. Download the key from Google Cloud IAM, paste it into Fivetran, and grant the service account BigQuery Data Editor and BigQuery Job User roles on your project.

Step 3: Add Your First Connector

Navigate to Connectors and click Add connector. Fivetran's connector catalog includes over 500 sources organized by category: databases, SaaS applications, file storage, event systems, and marketing platforms.

For a common connector like Salesforce:

1. Search for and select Salesforce
2. Authenticate with your Salesforce credentials (OAuth flow, no manual token handling)
3. Choose the Salesforce objects you want to sync: Accounts, Contacts, Opportunities, etc.
4. Set the sync frequency: options range from every 5 minutes to every 24 hours
5. Click Save and Test

Fivetran creates a schema in your destination named after the connector. All Salesforce objects sync into their own tables within that schema. An account object in Salesforce becomes a salesforce.account table in your warehouse.

For database connectors like PostgreSQL, the process involves enabling logical replication on your Postgres instance, creating a replication user with appropriate permissions, and providing the host, port, database, and credentials. Fivetran's documentation provides the exact PostgreSQL commands. Change data capture via logical replication means Fivetran tracks inserts, updates, and deletes in near-real-time rather than doing full table scans.

Step 4: Run and Verify Your First Sync

After setup, Fivetran initiates a historical sync that loads all existing data from the source. For large Salesforce orgs with hundreds of thousands of records, this initial sync can take 30 to 90 minutes. Subsequent syncs are incremental and typically complete in under 5 minutes.

To verify the sync:

1. Go to the connector dashboard and wait for the status indicator to show Synced
2. Navigate to your warehouse and query the synced tables
3. Check row counts against source system counts

A simple verification query in Snowflake:

SELECT COUNT(*) FROM salesforce.account;

Compare this against the record count in your Salesforce reports view. Within 5 to 10 percent is normal for the initial sync; exact parity depends on which objects and fields were selected.

Step 5: Manage Sync Health

Fivetran surfaces sync errors in the connector dashboard and sends email alerts when a sync fails. Common failure modes:

• API rate limit errors: Fivetran backs off and retries automatically, but high-frequency syncs on heavily used Salesforce orgs can hit limits. Reduce sync frequency or request a rate limit increase from Salesforce.
• Schema drift: When a source adds a new required field, Fivetran logs a warning and may pause the sync. Acknowledge the change in the dashboard to resume.
• Credential expiration: OAuth tokens for SaaS connectors expire periodically. Fivetran sends a warning email before expiration; re-authenticate through the connector settings.

Fivetran's Data Health tab shows sync latency, error rates, and row volume trends over the past 30 days. For teams running downstream dashboards, this is the first place to check when a report shows stale data.

Fivetran vs Self-Managed Alternatives

The main trade-off with Fivetran is cost versus control. Airbyte, the open-source alternative covered separately, requires you to host and maintain the sync infrastructure. For small engineering teams or non-technical operators, the operational overhead of self-hosting usually exceeds Fivetran's monthly cost at volumes under 10 million MAR.

Fivetran's connector reliability is its main advantage. Connectors are maintained by Fivetran's engineering team and updated when source APIs change. A self-managed Stripe connector breaks silently when Stripe deprecates an endpoint; Fivetran's breaks in a dashboard alert you can act on.

If you work with a dataset that is already loaded and just want to run analysis and get charts without setting up any pipeline at all, VSLZ lets you upload the file directly and ask questions in plain English.

What to Do After Your First Sync

Once data lands in your warehouse, the typical next step is transformation. Fivetran has a native dbt integration that lets you run dbt models on a schedule tied to connector syncs, so your cleaned and joined tables stay fresh automatically. Connect dbt Cloud to your destination and point Fivetran's transformation trigger at your dbt job ID.

For teams not using dbt, Fivetran's built-in transformations support basic SQL models that run post-sync. These work for simple renaming and type casting but are not a replacement for a full transformation layer.

With your warehouse populated and syncing, you can connect any BI tool, Looker Studio, Metabase, or Grafana, and point it at the destination schema.