What Does a Modern Data Team Look Like? Roles, Structure, and Scope

A modern data team is not a single monolithic function. It encompasses distinct roles with different technical skill sets, different stakeholders, and different accountabilities. Understanding who does what — and how these roles are typically organized — is prerequisite to building a data team that is productive rather than perpetually understaffed or misaligned.

The Core Roles

Data Engineer

Data engineers build and maintain the infrastructure that moves data from source systems to the places it needs to be for analysis. Their work includes: building and maintaining ingestion pipelines (Fivetran, Airbyte, custom CDC), designing and building data warehouse schemas, writing and maintaining dbt transformation models, orchestrating pipelines with Airflow or Prefect, managing data warehouse infrastructure and access, and building streaming pipelines when required.

Data engineers are infrastructure-first. Their primary concern is that data arrives where it needs to be, on time, with high reliability. They are typically strong in SQL, Python, and cloud infrastructure. They think in terms of pipelines, reliability, and scale.

Analytics Engineer

The analytics engineer (a relatively new role formalized by dbt) sits between data engineering and data analysis. They own the transformation layer: writing and maintaining the dbt models that transform raw ingested data into clean, business-logic-enriched tables that analysts and BI tools query. They define data models, write documentation, build data tests, and create the analytical foundation that other teams depend on.

Analytics engineers are SQL-first. They understand both the technical structure of the underlying data and the business logic that should be applied to it. They are the bridge between what the data warehouse contains and what business users need.

Data Analyst

Data analysts answer business questions using data. They write SQL queries, build dashboards and reports in Tableau or Power BI, produce ad-hoc analyses for stakeholders, and communicate findings to business teams. They are the most directly business-facing role in the data team.

Strong data analysts combine technical proficiency (SQL, data visualization, basic statistics) with domain knowledge and communication skills. They understand what the business stakeholder is actually trying to learn, not just what they asked for.

Data Scientist

Data scientists build predictive and statistical models. They work on problems that go beyond descriptive analytics: customer churn prediction, demand forecasting, fraud scoring, recommendation systems, NLP on text data, A/B test analysis. Data scientists use Python or R, work with ML frameworks (scikit-learn, XGBoost, PyTorch), and require clean feature data from the warehouse.

In most mid-market organizations, data science is a smaller function than analytics or engineering. Many organizations benefit more from improving their descriptive and diagnostic analytics than from building complex ML models.

BI Developer / BI Engineer

BI developers specialize in building and maintaining BI tool environments: Tableau Server, Tableau Cloud, Power BI Premium, Looker deployments. Their work includes: designing and building dashboards and reports to production quality, managing Tableau Server administration (extracts, permissions, performance), building certified data sources, and maintaining governance standards for published content.

In Tableau-centric environments, the BI developer role may be the most critical data team function — responsible for the analytical environment that business users interact with daily.

Data Manager / Director of Data

The data leadership role defines the data team's strategy, manages stakeholder relationships, prioritizes the roadmap, and ensures the team is aligned with business objectives. They translate between technical capabilities and business needs, manage headcount and budget, and represent the data function in cross-functional discussions.

How Data Teams Are Organized

**Centralized team** — all data roles report to a central data function (typically a VP of Data, Head of Data, or CTO). The team serves all business domains from a central function. Simple to manage and enables cross-domain data consistency. Risk: the team becomes a bottleneck, with all domain requests funneled through a single team.

**Embedded team** — data roles are distributed into business units. The sales team has its own analyst; the finance team has its own analyst. Deeply aligned with domain needs. Risk: inconsistent methodologies, duplicated infrastructure, siloed data, incompatible metric definitions.

**Hub-and-spoke model** — a central team owns shared infrastructure (data engineering, core warehouse models, BI platform), with embedded analysts or analytics engineers in each domain. The hub ensures consistency and infrastructure quality; the spokes provide domain alignment. The most common model for mid-to-large organizations.

Team Sizing and Evolution

For organizations just building out a data function:

**1–3 people**: typically a generalist data analyst/engineer who does a bit of everything — queries the database, builds reports in Tableau, does ad-hoc analysis. No formal data engineering. Limited pipeline automation.

**4–8 people**: beginning to specialize. One or two dedicated analytics engineers building the warehouse layer. Two or three analysts serving different domains. A BI developer managing the Tableau or Power BI environment. Light data engineering.

**8–20 people**: meaningful specialization. Data engineering team building and maintaining pipelines. Analytics engineering team owning the dbt layer. Analytics team serving multiple domains. Data science function. BI platform team. Beginning to need a manager or director.

**20+ people**: full discipline specialization. Platform engineering separate from application data engineering. Multiple analytics teams by domain. Data science with ML engineering. Strong governance function. Data leadership at the director or VP level.

The right team structure for any organization depends on the data maturity, the scale of the data environment, and the business's dependence on data for operations and strategy. Most organizations understaff their data teams relative to the expected output — a persistent mismatch that produces perpetual backlog.

Our data architecture and BI strategy practices help organizations design data team structures and operating models — contact us to discuss your data team design requirements.

A former Microsoft data architect audits your data foundation, identifies your top priorities, and sends you a written plan. Free. No pitch.