ELT (Extract, Load, Transform) is the modern data integration pattern enabled by cloud data warehouses. Data is extracted from source systems (like NetSuite), loaded raw into a destination warehouse (Snowflake, BigQuery, Redshift, Synapse), and then transformed using the warehouse's compute. ELT inverts the traditional ETL order, which transformed data before loading. ELT became viable with the rise of cheap, scalable cloud warehouses where compute is decoupled from storage.
ELT vs ETL — order matters
| Aspect | ETL | ELT |
|---|---|---|
| Order | Extract → Transform → Load | Extract → Load → Transform |
| Transform happens | In flight, in ETL server | In destination warehouse |
| Raw data preserved? | No | Yes (raw stays) |
| Setup time | Slow (define schema upfront) | Fast (load first, model later) |
| Debug | Hard (in-flight) | Easy (re-query raw) |
| Cost at scale | ETL server compute | Warehouse compute (often cheaper) |
| Best for | Sanitized data into traditional DBs | Cloud-warehouse analytics |
ELT for NetSuite
For NetSuite analytics, ELT is the dominant pattern:
- Extract via SuiteQL through NetSuite REST Web Services
- Load raw tables into Snowflake / BigQuery / SQL Server / Azure Synapse / Redshift
- Transform in warehouse: star schema, currency conversion, OneWorld consolidation, KPI calculations
Acterys NetSuite Sync follows this pattern. SuiteQL pulls raw NetSuite data into your warehouse, and Acterys applies a pre-built star schema transformation layer optimized for Power BI. Custom transformations can be added downstream via dbt or warehouse SQL.
When ETL still wins
- Regulated industries — data must be sanitized/masked before landing
- Traditional row-oriented destinations — Oracle DB, SQL Server on-prem, MySQL
- Small data volumes — in-flight transformation cheaper than warehouse compute
- Network constraints — limited bandwidth to cloud warehouse
Common ELT tooling for NetSuite
- Acterys NetSuite Sync — SuiteQL native, pre-built star schema, no ODBC
- Fivetran — generic ELT, requires SuiteAnalytics Connect (ODBC), MAR pricing — see MAR pricing glossary
- Stitch — similar to Fivetran, less NetSuite-specific tuning
- Airbyte — open-source ELT, NetSuite connector available
- Custom (Python + REST + dbt) — full control, ongoing dev cost
Related glossary entries
- SuiteQL — the "E" in ELT for NetSuite
- Incremental sync — efficient "E" for delta-only fetches
- MAR pricing — how Fivetran charges for ELT