dbt in 2026: Why Analytics Engineers Are the Most Valuable People on Your Data Team

Analytics Engineering Has Grown Up

When dbt first became mainstream around 2020–2021, analytics engineering felt like a transitional role — halfway between data analyst and data engineer. Companies weren't quite sure where to put it on the org chart. Salaries were all over the place.

In 2026, that ambiguity is largely gone. Analytics engineering is a defined discipline with clear expectations, a strong community, and compensation that regularly outpaces pure analytics.

What the Role Actually Looks Like

An analytics engineer owns the transformation layer. You're taking raw data that lives in your warehouse (Snowflake, BigQuery, Databricks — pick your flavour) and turning it into models that analysts, data scientists, and product teams can actually use.

In practice this means:

• Writing and maintaining dbt models
• Designing data marts and semantic layers
• Working closely with stakeholders to understand how data is consumed
• Building data quality tests that actually catch issues before they reach dashboards

The dbt Skills That Separate Seniors from Juniors

Most people who say they "know dbt" know how to write a basic model and run 'dbt run'. That's table stakes.

What separates senior analytics engineers:

Modular design: Knowing when to use staging vs intermediate vs mart models. Understanding the medallion architecture not just conceptually but in practice for specific team needs.

Macros and packages: Writing reusable macros that your team actually uses. Knowing which dbt packages are worth adopting and which add unnecessary complexity.

Testing philosophy: Going beyond 'not_null' and 'unique' to write custom tests that reflect actual business logic. Knowing which tests are worth the maintenance cost.

Performance thinking: Understanding how your models translate to warehouse query costs. Materialisation decisions (table, view, incremental) that reflect real usage patterns.

Where the Market Is Going

The most interesting trend I'm seeing: analytics engineering is absorbing more of what used to live in the data engineering world. As tools like dbt and the modern data stack mature, the line between "transformation" and "orchestration" is blurring.

Analytics engineers who understand Airflow basics, can write a bit of Python, and can speak to data infrastructure decisions will command significant premiums in 2026.

Getting There

If you're an analyst looking to transition: start with dbt Fundamentals (it's free), then build something real — take a dataset from your current work and build a proper dbt project on top of it. The portfolio signal is strong.

If you're already doing analytics engineering but want to level up: go deeper on the infrastructure side. Learn how your warehouse is actually processing your queries. That knowledge will make you dramatically more effective.