Most organisations believe they have a data capability.

In reality, they have a collection of job titles with overlapping responsibilities, unclear ownership, and no consistent path from data to decision. This is the root cause of slow, expensive, and underwhelming data projects. The issue is not talent. It is design.

The Core Mistake: Confusing Job Titles with Work

Terms like “Data Analyst”, “Data Scientist”, “BI”, and “Data Governance” are treated as clearly defined roles.

They are not.

They are capabilities that exist at different stages of a data workflow.

When organisations design teams around these titles instead of the underlying work, three things happen:

• Critical responsibilities are missed
• Other responsibilities are duplicated
• Accountability becomes unclear

The result is friction at every stage of delivery.

The 6 Jobs That Exist in Every Data Workflow

Regardless of industry or scale, every data initiative requires six distinct jobs to be done.

1. Data Creation

Job: Capture data at the source.

This responsibility typically sits in business operations, front office teams, and core systems.

This is where data quality is either created or destroyed.

Common failure: Speed is prioritised over accuracy.

Impact: Poor data quality propagates downstream, increasing cost and complexity.

2. Data Governance

Job: Define and enforce what “good data” means.

This includes standards, definitions, ownership, and controls.

Common failure: Governance exists as documentation, not enforcement.

Impact: Teams agree on definitions in theory, but diverge during delivery.

3. Data Engineering

Job: Move, transform, and structure data.

Pipelines, integration, and data models sit here.

Common failure: Optimisation for architecture over usability.

Impact: Data is technically correct but difficult to use or trust.

4. Data Analysis

Job: Turn inconsistent data into decision-ready information.

This is the core of value creation in most data projects.

Common failure: Analysts are pulled into reporting, engineering, or administrative work.

Impact: The actual business problem remains unsolved.

5. Business Intelligence (BI)

Job: Distribute and scale access to information.

Dashboards, reporting layers, and self-service tools sit here.

Common failure: Focus on output (charts) instead of outcomes (decisions).

Impact: Increased reporting volume without improved decision-making.

6. Data Science

Job: Predict, optimise, or automate decisions.

This layer builds on the foundations created by the previous stages.

Common failure: Used prematurely as a shortcut.

Impact: Advanced models built on unstable data foundations.

The Missing Layer: Decision Ownership

None of the above creates value unless decisions are made.

Job: Act on insights.

This responsibility sits with the business, not the data team.

Common failure: Decision-making is delegated back to data teams.

Impact: Endless analysis cycles with no measurable impact.

Why Most Data Teams Fail

Most organisations attempt to compress multiple (or all) of these responsibilities into a single role:

👉 “Data Analyst”

This leads to:

• Burnout
• Poor governance adoption
• Overproduction of dashboards
• Underperformance of data science initiatives

The system is misconfigured from the start.

The Implication

Improving tools, hiring more people, or adding AI will not fix this.

If the underlying role design is wrong, improvements in capability will only accelerate inefficiency.

What High-Performing Teams Do Differently

They do one of two things well:

1. Separate responsibilities clearly across the workflow
2. Deliberately automate specific layers while maintaining ownership clarity

They do not rely on ambiguous, overloaded job titles.

Final Thought

If you are hiring for a “Data Analyst”, you are making a design decision whether you realise it or not.

The critical question is: Which of these jobs are you actually asking them to do?

If the answer is “most of them”, the problem is not the individual. The problem is the system they are being asked to operate within. If you want to design a data team that actually delivers outcomes, start by redesigning the roles around the work — not the titles.