What is Azure Data Factory? A plain-English explainer

What ETL and data pipelines are

Most organisations hold data in more places than one system can use directly — a transactional database, a set of application logs, a third-party API, a spreadsheet someone maintains by hand. Getting that data into a usable, combined shape usually means extracting it from its source, transforming it (cleaning, reshaping, combining), and loading it into a destination built for analysis. Our what-is-etl explainer covers that pattern in depth, and what-is-a-data-pipeline covers the broader idea of an automated, repeatable flow of data from one place to another.

Data Factory exists to build and run exactly that kind of pipeline, without you writing and scheduling the movement logic entirely by hand.

What Data Factory actually does

Data Factory is built around pipelines — defined sequences of steps that move and transform data — which you can build visually or through code, and then schedule or trigger to run.

• Ingestion — connects to a wide range of data sources, from databases and file stores to software-as-a-service applications.
• Transformation — reshapes, cleans or combines data as it moves, using either visual data flows or external compute services.
• Orchestration — sequences and schedules multi-step data workflows, including dependencies between steps.
• Monitoring — tracks pipeline runs, so failures and delays are visible rather than silent.

Where Data Factory fits, and Microsoft Fabric

Data Factory typically sits upstream of the systems that actually analyse the data — feeding a data lake, a data warehouse, or a reporting layer built on top of one. Our data-lake-vs-data-warehouse explainer covers the difference between those two destinations; Data Factory is largely indifferent to which one you are feeding, because its job is the movement and transformation, not the final storage model.

Worth knowing honestly: Microsoft Fabric is Microsoft’s newer, unified data platform, and it brings data-integration capabilities in the spirit of Data Factory together with data warehousing, analytics and other data workloads under one product. Fabric does not erase Data Factory — Data Factory remains a standalone Azure service — but if you are starting fresh on Microsoft’s data stack today, Fabric is worth knowing about as the direction the platform is heading.

The Azure equivalent of Glue and Dataflow

Every major cloud provider offers a comparable data-integration service. Azure Data Factory is broadly the counterpart to AWS Glue and Google Cloud Dataflow — all three exist to move and transform data between sources and destinations at scale, without you building that orchestration entirely from scratch. The specifics of how each one is built and priced differ, but the underlying job — connect to data, transform it, move it reliably and on a schedule — is the same.

How you pay for it, without the numbers

Data Factory follows the consumption-based pricing common across data-integration services: you are billed based on how much your pipelines actually run and process, rather than a fixed cost for having the service available. Exact pricing units, rates and free allowances change over time, so this article deliberately avoids listing them — check Microsoft’s official Data Factory pricing page for current detail.

Where Data Factory appears in certification study

Data Factory is core to Azure’s data-engineering path. It is directly relevant to certifications focused on building and managing data platforms on Microsoft’s stack, including those covering Microsoft Fabric, where data-integration concepts carry over even as the tooling evolves. It also appears, at a conceptual level, in the broader Azure Administrator syllabus.

This article covers what Data Factory is and why it exists — pipeline design, connectors and transformation activities belong in the syllabus itself, which our /revision study library breaks down lesson by lesson for these certifications.