Home · Free tools · Glossary · ETL
What is ETL?
ETL stands for extract, transform, load: pull data out of a source system, reshape it, and put it into a destination — usually a warehouse — on a schedule. It is the plumbing under most reporting that does not involve a person exporting a CSV.
Why the letters got reordered
ETL's order exists because storage and compute used to be expensive: you transformed data before loading it, so you only stored what you needed. When warehouses became cheap and fast, the order flipped to ELT — load the raw data first, transform inside the warehouse afterwards.
The practical difference: with ELT you still have the raw data when someone asks a question your transform did not anticipate. With ETL, that data was discarded on the way in. This is why ELT won for analytics, and why ETL persists where the destination is an operational system rather than a warehouse.
The T is where the work is
Extract and load are mostly solved by connectors. Transform is where the judgment lives: deduplicating, reconciling identifiers that disagree between systems, normalizing currencies and time zones, deciding what "active customer" means. That work does not get easier with better tooling, because it is not a tooling problem.
When you do not need any of this
ETL is infrastructure, and infrastructure has a floor cost — a schedule to maintain, failures to notice, a warehouse to pay for. A great deal of real reporting is one person, a few sources and a monthly cadence. That case is served by something smaller: the same steps kept next to the files and run again on next month's exports. TableDI 2 is a desktop app for file work you redo every period: it keeps your sources, rules and delivery as a job, so next month you drop in the new files and run it again. Everything runs on your own machine.
The line is roughly: if several people depend on the same definitions and the data is bigger than one machine, build the pipeline. If it is you, a laptop and a monthly report, do not.
Questions people ask
What is the difference between ETL and ELT?
The order of the last two steps. ETL transforms before loading, so only the transformed data is stored. ELT loads raw data first and transforms inside the warehouse, which keeps the raw data available for questions the original transform did not anticipate.
Do I need ETL for a monthly report?
Usually not. A pipeline has a floor cost in maintenance and infrastructure. One person, a few sources and a monthly cadence is better served by steps that live next to the files and run again each month.
Is a spreadsheet an ETL tool?
Not in the infrastructure sense — no scheduling guarantees, no lineage, no orchestration. But the extract-transform-load shape is what a monthly spreadsheet routine does by hand: pull the exports, clean them, and paste the result into the report.
Related
Doing this every month?
In TableDI 2 you do it once, then save it as a job. Next month you drop in the new files and run it again.
macOS, Apple silicon and Intel; Windows is in progress (what to do meanwhile). Free is not a trial — no account, no card.
Last reviewed 2026-09-11 by the TableDI team. Something wrong on this page? Tell us — it is one inbox, read by the people who build TableDI.