What is Data Transformation?
The process of converting data from one format, structure, or value system to another, making it suitable for analysis, reporting, or use in downstream systems.
Understanding the Details
Data transformation is where raw data becomes useful. It includes cleaning (removing errors), standardising (normalising formats), enriching (adding context), aggregating (combining records), and restructuring (reshaping for different uses). In an ETL/ELT context, transformation happens either before loading (ETL) or after loading into the warehouse (ELT). Common transformations include converting time zones, mapping product codes to names, calculating derived metrics (like MRR from billing data), joining data from multiple sources, and filtering irrelevant records. dbt has become the standard tool for transformation in modern data stacks, enabling version-controlled, tested, and documented transformation logic.
How It Works in Practice
Revenue calculation
Raw Stripe events are transformed into MRR calculations by handling upgrades, downgrades, churn, and expansion across billing periods.
Data standardisation
Contact records from different sources are transformed to use consistent date formats, phone number patterns, and address structures.
Metric derivation
Raw event data is transformed into calculated metrics: session duration, pages per session, conversion rates, and engagement scores.
Why It Matters
Raw data is rarely in the format needed for decisions. Transformation is what turns scattered, inconsistent data into the clean, structured information teams can act on.
What People Often Get Wrong
Transformation is a technical task with no business input needed. Actually, business context is essential for correct transformation logic and metric definitions.
Once transformation logic is built, it runs forever. Actually, transformations need updating as source schemas change and business definitions evolve.
All transformation should happen in one place. Actually, different transformation types may happen at different stages of the data pipeline.
How We Handle Data Transformation
We build transformation layers with clear documentation, testing, and version control so your data models are reliable and maintainable as requirements evolve.
Related Terms
Common Questions
Need Help With Data Transformation?
If you'd like to discuss how data transformation applies to your business, we're happy to explain further.