Data
ETL / ELT
Also called
- Data pipeline
- Extract transform load
ETL transforms data before loading it; ELT loads it raw and transforms it in the warehouse. ELT has become the default because storage is cheap and keeping the raw copy means a modeling mistake is fixable without re-extracting.
Pipelines fail constantly and quietly. A source changes a field name, an API deprecates a version, a file lands late. A pipeline without monitoring does not stop — it produces yesterday's numbers indefinitely, which is worse than producing none.
Freshness expectations should be explicit and matched to the decision. Hourly costs meaningfully more than daily, and most decisions do not need it.
Silent pipeline failure is the most common cause of a business acting on stale data without knowing it.
Commonly misunderstood
What people get wrong
The claim
“We set the pipeline up, so the data is current.”
What is actually true
Only if something checks. Freshness and row-count alerting is a small amount of work that catches the failure mode that costs the most.
Where this comes up
Services where it matters
Related terms
Next step
Working through a etl / elt decision?
Tell us the situation. We will give you the tradeoffs as we see them, including when the answer is that you do not need what you are being sold.
No pitch deck. A 30-minute conversation about what you are trying to achieve.