Glossary
What is the difference between ETL and reverse ETL?
ETL moves data from operational systems into a data warehouse; reverse ETL moves modelled data from the warehouse back into operational tools.
ETL (extract, transform, load) copies data from operational systems such as an HRIS, CRM or LMS into a data warehouse for reporting and analysis. Reverse ETL does the opposite: it takes cleaned, modelled data from the warehouse and syncs it back into operational tools, so teams can act on it where they work.
How does it work?
| ETL / ELT | Reverse ETL | |
|---|---|---|
| Direction | Apps → warehouse | Warehouse → apps |
| Purpose | Reporting, analytics | Activation, operations |
| Typical target | Snowflake, BigQuery | CRM, LMS, support tools |
| Typical tools | Fivetran, Airbyte | Hightouch, Census (now part of Fivetran) |
Many teams now load raw data first and transform it inside the warehouse, a pattern called ELT. The direction is the same as ETL; only the order of steps changes.
Where it shows up in integrations
- Learning analytics LMS completions flow into the warehouse via ETL to combine with HR data.
- Customer health A score computed in the warehouse is pushed to the CRM via reverse ETL.
- Compliance reporting Training status by department is modelled in the warehouse and sent back to managers’ tools.
Common pitfalls
- Using reverse ETL as a replacement for real-time integration when users need immediate updates.
- Overwriting fields in operational tools that people also edit by hand.
- Not defining which system owns each field, which causes sync loops.