Before analytics engineering existed as a distinct role, data teams had a persistent problem: data engineers built pipelines with software engineering rigor but without deep business context, and business analysts and data analysts had strong business context but operated on data they could not trust because it was inconsistently cleaned and transformed. The analytics engineer sits in between, owning the data transformation layer — taking raw data from the warehouse and building clean, tested, documented models that analysts can use with confidence. This structural role is now recognized as genuinely distinct and is one of the fastest-growing data roles by job posting volume.
Why dbt specifically created the role
dbt operationalized the workflow: writing SQL transformations as code, running automated tests on data assumptions, generating documentation from the code, and managing dependencies between models through a DAG. Before dbt, data transformation was done in a mix of stored procedures, Python scripts, and manually maintained spreadsheets. dbt applied software engineering practices to what had been an ad-hoc discipline, which created a coherent enough workflow that it could be hired for specifically. The role did not fully exist before the tooling existed to define it — which is unusual in the data industry and explains why analytics engineering feels both new and immediately legible to anyone who has worked in data.
The salary premium relative to analyst roles
Analytics engineers at mid-sized tech companies earn $100k-$160k. Senior analytics engineers at major tech companies earn $150k-$200k. This compares favorably to data analyst roles at $80k-$130k for similar experience levels. The premium reflects the combination of SQL depth and software engineering practice that the role requires. Analysts who want to move into analytics engineering need to learn dbt, become comfortable with version control in git, and develop a habit of writing tests on their transformations — not as an afterthought but as a core part of the workflow. The practical entry point is the dbt Learn curriculum, which is free and takes about five hours to complete, followed by a small portfolio project that demonstrates models, tests, and documentation together.