Managed data
What is Data as a Service (DaaS) and why it matters
A clear definition of Data as a Service (DaaS): what it includes, how it differs from building your own infrastructure and why more companies adopt it.
Read articleKey takeaways
- ETL transforms data before loading; ELT loads first and transforms in the destination.
- ELT leverages modern storage power and scales better with large volumes.
- ETL is still useful for complex rules or pre-load compliance.
- The choice depends on volume, infrastructure and data sensitivity.
- Many architectures combine both.
ETL and ELT describe two ways of moving and preparing data. The difference is a single letter — the order of transformation — but it has real consequences for performance, cost and governance.
The difference
ETL (Extract, Transform, Load) extracts data, transforms it in an intermediate environment and loads it ready. ELT (Extract, Load, Transform) loads raw data into the destination and transforms it there.
ETL
Transformthen load
ELT
Loadthen transform
Choose by
Volume, infrasensitivity
Why ELT has gained ground
Scalable storage and compute have shifted the balance to ELT. Loading first and transforming in the destination leverages its processing power, keeps raw data available for new uses and scales with volumes that would overwhelm an intermediate environment.
When ETL still makes sense
ETL is preferable when transformations are very complex, or when quality or anonymisation rules must be applied before data lands — for example with sensitive personal data, where transforming before loading can be a privacy requirement.
How to choose
- Volume: more data favours ELT.
- Infrastructure: a modern warehouse or lake enables ELT.
- Sensitivity: may require prior transformation (ETL).
- Combine: ELT for most flows, ETL for special cases.
It is not ETL or ELT: most architectures combine both depending on the flow.
In summary
ETL transforms before loading; ELT loads first and transforms in the destination, scaling better with large volumes and modern cloud. ETL remains preferable for complex rules or pre-load privacy. The choice depends on volume, infrastructure and sensitivity — and many architectures combine both.
Sources & further reading
Frequently asked questions
Is ELT always better than ETL?
No. ELT scales better with large volumes and modern architectures, but ETL is preferable for complex rules or privacy requirements that demand transforming before loading.
Do I need to choose only one?
No. It is common to combine both: ELT for most flows and ETL for cases with complex transformations or sensitive data.
What about data quality?
In both, applying validation and cleaning is key — before loading in ETL, or via governed transformations in the destination in ELT.
Why has ELT become popular?
Because scalable storage and compute let you transform in the destination, leveraging its power and keeping raw data available for new uses.
When is ETL the right choice?
For very complex transformations, or when quality or anonymisation rules must be applied before data lands in the destination.
What determines the choice?
Volume (favours ELT), destination infrastructure (enables ELT) and data sensitivity (may require ETL).
Turn this data into results
Tell us what you want to achieve. Data Layer connects, processes and delivers the result up and running, with no infrastructure for you to manage.