Census
Overview
Census is the leading Reverse ETL platform that syncs data from your data warehouse to the business tools your teams use every day. Founded in 2018, Census has pioneered the "operational analytics" category, enabling data teams to activate their warehouse data in sales, marketing, and success tools without writing code.
What is Census?
Census is a Reverse ETL platform that operationalizes your data warehouse by syncing clean, modeled data to 200+ business applications. Unlike traditional ETL (Extract, Transform, Load data into warehouse), Reverse ETL moves data from the warehouse to where business teams need it.
Core Capabilities:
- No-Code Syncs: Configure data flows through UI, no custom scripts
- Real-Time & Scheduled: Sync on schedules or trigger-based
- Bi-Directional: Read and write to/from SaaS tools
- Data Quality: Built-in validation and error handling
- Warehouse-Native: Uses your warehouse as single source of truth
Why Use Census?
The Problem Census Solves
Before Census:
With Census:
Key Benefits
- Warehouse as Source of Truth: Leverage existing dbt models, no duplicate logic
- Self-Service for Business Teams: Non-technical users configure syncs
- No Custom Code: Visual mapping reduces engineering burden
- Scalability: Handle millions of records without performance issues
- Data Governance: Centralized visibility and control
When to Use Census
- Sales & Marketing Ops: Sync lead scores, customer segments, product usage to CRM
- Customer Success: Push churn risk, health scores to support tools
- Product Analytics: Sync feature usage, engagement metrics to marketing tools
- Finance: Update ARR, LTV, payment status in business systems
- Data Democratization: Enable teams to access warehouse data in familiar tools
Core Concepts
Sources
Your data warehouse where clean, modeled data lives:
Supported Sources:
- Snowflake (most common)
- BigQuery
- Redshift
- Databricks
- PostgreSQL
- MySQL
- Synapse
- Rockset
Best Practice: Use dbt to create mart tables specifically for Census syncs
Destinations
Business tools where you want to send data (200+ supported):
Popular Destinations:
- CRMs: Salesforce, HubSpot, Pipedrive, Copper
- Marketing: Braze, Iterable, Customer.io, Marketo, Mailchimp
- Support: Zendesk, Intercom, Front, Gorgias
- Ad Platforms: Google Ads, Facebook Ads, LinkedIn Ads
- Product Analytics: Amplitude, Mixpanel, Heap
- Data Apps: Slack, Notion, Airtable, Google Sheets
Syncs
A sync defines how data flows from source to destination:
Sync Configuration:
Sync Modes
Update or Create (Upsert):
- Create records if they don't exist
- Update existing records
- Most common mode
Update Only:
- Only update existing records
- Don't create new ones
- Useful for enrichment
Mirror:
- Keep destination in perfect sync with source
- Deletes records not in source
- Use carefully!
Append:
- Only insert new records
- Never update or delete
- Good for event logs
Sync Keys
Unique identifiers to match source and destination records:
Examples:
Composite Keys (multiple fields):
Field Mappings
Map source columns to destination fields:
Direct Mapping:
Transformations (Limited):
- Census intentionally limits transformations
- "Transform in warehouse, not in Census"
- Some basic formatting available (dates, booleans)
Lookups:
Conditional Syncing:
Architecture
How Census Works
Execution Flow:
- Census queries warehouse for source data
- Identifies changed rows (diffing)
- Batches API calls to destination
- Handles rate limits and retries
- Logs successes/failures
- Notifies on errors
Security
Warehouse Access:
Destination Access:
- OAuth connections (recommended)
- API keys (where OAuth unavailable)
- Credentials encrypted at rest
- Audit logs for all data access
Common Workflows
Sales Enablement
Sync Product Usage to Salesforce:
Census Sync:
Result: Sales reps see real-time product usage in Salesforce, identify expansion opportunities
Marketing Automation
Sync High-Value Leads to Ad Platforms:
Census Sync:
Result: Target lookalike audiences based on data warehouse predictive models
Customer Success
Sync Churn Risk to Intercom:
Census Sync:
Result: CS team receives automated alerts for at-risk customers
Census + dbt Integration
The Golden Workflow:
Exposures in dbt (Track Census dependencies):
Benefits:
- dbt lineage shows which models power Census syncs
- Impact analysis before changing models
- Documentation of downstream dependencies
Advanced Features
Audiences (Segments)
Create reusable user segments:
Entity Resolution
Match records across systems:
Custom Objects
Sync to non-standard destination objects:
Lookups
Resolve foreign keys automatically:
Computed Columns
Light transformations in Census (use sparingly):
Best Practice: Do transformations in warehouse/dbt instead
Monitoring & Observability
Sync Monitoring
Census Dashboard shows:
- Last Sync Status: Success/Failed/In Progress
- Records Processed: Created, Updated, Failed
- Sync Duration: Historical performance
- Error Details: Why specific records failed
Alerts
Configure notifications:
- Sync failures
- High error rate (> 5%)
- Sync duration anomalies
- Validation failures
Channels:
- Slack
- PagerDuty
- Webhook (custom integrations)
Logs
Detailed logs for troubleshooting:
Export Logs: Download CSV of failed records for remediation
Pricing
Census pricing based on:
- Number of Synced Rows (per month)
- Number of Destinations
- Sync Frequency (real-time costs more)
Tiers (approximate):
- Core: $500-$1,000/month (basic syncs, scheduled)
- Growth: $2,000-$5,000/month (more rows, faster syncs)
- Enterprise: Custom pricing (real-time, high volume, premium support)
Free Trial: 14 days, full access
Alternatives
| Platform | Focus | Best For |
|---|---|---|
| Census | Reverse ETL | Warehouse-first teams, dbt users |
| Hightouch | Reverse ETL | Similar to Census, strong competitor |
| Fivetran | Forward ETL | Getting data into warehouse (opposite direction) |
| Segment | CDP | Event tracking, customer data platform |
| Polytomic | Reverse ETL | Smaller teams, cost-sensitive |
Census vs Hightouch:
- Very similar products
- Both excellent choices
- Census slightly better dbt integration
- Hightouch slightly better visual audience builder
When Not to Use Census
- Real-Time Event Streaming: Use Kafka/Kinesis instead
- Two-Way Sync Critical: Census is primarily one-way (warehouse → tools)
- No Data Warehouse: Census requires warehouse as source of truth
- Very Low Volume: Custom scripts might be simpler for 1-2 syncs
- Complex Transformations Needed: Do in dbt/warehouse first
Getting Help
- Documentation: https://docs.getcensus.com/
- Community Slack: https://www.getcensus.com/community
- Support: support@getcensus.com (or in-app chat)
- Office Hours: Weekly Q&A sessions (Enterprise customers)
- GitHub Issues: Public feature requests
Ready to activate your warehouse data? Check out: