jemoraz

Building a Power BI Design System

The idea of creating a design system for Power BI didn’t come out of nowhere.

It began at PepsiCo, where I was working with the design team on systems for web and mobile applications. When we tried to apply those same guidelines to Power BI, it just didn’t work: what made sense in traditional digital interfaces wasn’t viable in BI environments.

That mismatch pushed me to dig deeper. I realized that designing data products in Power BI required a different approach — with rules adapted to its context: from how colors behave in charts, to how templates and accessibility are managed in a tool built more for analysts than designers.

That initial curiosity became the starting point for a bigger project: building a dedicated Design System for Power BI.

STEP 1

The problem I found

When I stepped into this new challenge, it quickly became clear that the dashboard ecosystem was fragmented.

Each domain had its own style, KPIs were named differently, and makeshift templates were rarely reused. For users, moving from one report to another felt like switching languages — confusion, low adoption, and wasted time.

Overloaded with colors and icons, the dashboard lacks consistency and structure, making data hard to read and compare.

Weak color hierarchy and minimal labels make comparisons unclear, while the layout feels fragmented and lacks narrative flow.

"I spent more time trying to understand the dashboard than actually using the data to take action."

Operations stakeholder

19%

Clarity

Users struggled to interpret KPIs and follow data narratives.

25%

Consistency

Inconsistent design patterns across domains.

38%

Scalability

38% of misalignment stemmed from templates that couldn’t scale.

19%

Accessibility

Impact on limiting adoption across wider audiences.

Note: % of misalignment issues detected by principle across audited dashboards.

I realized that while data teams were highly focused on tools to speed up processing, cleaning, and reporting, what was missing was a product mindset — the kind digital product teams rely on.
Bringing in a UX perspective meant working side by side with Data Scientists, Analysts, Governance, and Business Owners to connect their needs into a shared vision.
To do this, I ran formal and informal interviews, mapping recurring pain points and discussing ideas together.
This became the foundation for uncovering the problems that the design system would later address.

STEP 2

Uncovering the problems

I first needed to understand where the chaos was coming from.

To answer that question, I first needed to understand where the chaos was coming from. The first step was to listen and map the terrain: I reviewed over 50 dashboards, interviewed analysts and business owners, and identified repeating issues such as poor contrast, inconsistent naming, and the lack of visual standards.
The treemap below shows the main issues I uncovered, each weighted by its impact on dashboard usability and adoption. ↓

STEP 3

The result

From the fragmented dashboards we uncovered — overloaded visuals, inconsistent naming, and the absence of visual standards — we moved to a system with clear, easy-to-apply rules.

In just a few weeks, the change was visible: the time needed to create a dashboard dropped from two weeks to just four days. Visual consistency improved by more than 70%, making reports easier to read and compare. Most importantly, stakeholders began to trust the dashboards again, because the information was finally presented in a way they could understand effortlessly.

Power BI design system deliverables

To make the results sustainable, the design rules were translated into tangible deliverables that both designers and analysts could use in their daily work.

Structured JSON theme files for consistency, a .PBIX reference file to showcase layouts and chart patterns, a Figma UI Kit for prototyping, and a custom icon library.

These deliverables became a single source of truth, ensuring smooth hand-off and scalable adoption across teams.

"Now we can deliver dashboards in days, not weeks. It changed how we trust our data."

Data analyst

STEP 4

The impact

A Power BI design system isn’t “extra decoration.” It’s a strategic tool that multiplies impact: it speeds up dashboard creation, improves adoption, and builds trust in the data.

This project confirmed for me that in the world of data products, design needs its own approach to truly be useful and scalable.

Dashboard creation

4 days

vs 14 days

Reducing design and setup time meant dashboards could be produced 3x faster.

Adoption

80%

vs low adoption

Most teams now rely on templates and guidelines as their default starting point.

Onboarding

7 days

vs 30 days

New analysts adopted the design system in just one week, cutting the learning curve by 75%.