A dashboard guide for teams without a designer
A reference guide for internal teams building dashboards without a designer, kept current for four years.
Context
Senior UX Designer, horizontal UX for an internal data org · 2020–2024
Services: UX consulting · Information architecture · Data visualization guidance · Content design · Office hours
The guide was internal, so the figures on this page are recreated with placeholder content. The principles are real; the layouts are not.
My org built dozens of internal dashboards and data views, and most of those teams had no designer. They asked for help one at a time, which didn't scale, and the answers they needed were scattered across several design systems, charting libraries and years of research nobody could find.
I took on consulting engagements across the org, from quick IA reviews to full product design for teams like payments reporting and planning tools. I built a framework for deciding which requests to take on and at what depth, wrote a short UX brief template so teams could describe their problem before we met, and held open office hours for everyone else.
The same questions kept coming up: what goes at the top, how much detail belongs on one screen, which chart and which component. So I wrote the answers down once.
The guide
A 34-page reference built around three ideas:
- Put what needs attention first. The highest-priority alert or action leads, then the key chart and the item it affects, then status and context.
- Match detail to scope. A dashboard that spans several services stays high-level, with few actions and links out. A single-service dashboard can go deeper and carry more actions.
- Don't rebuild what exists. Point teams to the design-system components, charting guidance, accessibility standards for data visualization and past research that already answered their question.
How it was organized
I ordered the guide the way a team builds a dashboard: write the brief, decide what goes first, decide how much belongs on the screen, choose the charts, pick the components, check accessibility, then look at what other teams had done. Few people read it front to back. Most arrived with one question from office hours or a design review, so every section had to work as a starting point on its own.
Each section followed the same pattern: the question as teams asked it, a short answer, an example, and links to the component, standard or research behind the answer. Once someone learned where to look on one page, they could find it on the next. The repeated structure also made the guide easier to keep current, because an update to one component or standard only touched the pages that linked to it.
What goes first
The most common problem was a dashboard that showed everything at the same weight, so nothing stood out. The guide broke a dashboard into three bands, read from the top:
- Needs attention. The single highest-priority alert or action, with one clear next step.
- The key chart and what it affects. The chart that explains the alert, next to the item it affects, so people don't have to cross-reference two screens.
- Status and context. Summary tiles, when the data was last refreshed and where it comes from.
1 · Needs attention
2 · Key chart and what it affects
3 · Status and context
How much belongs on one screen
Scope decided depth. A dashboard spanning several services is read by people checking the health of many things at once, so it stays high-level: a few summary numbers, one simple chart, and a link from each tile to the team that owns the detail. A dashboard for one service is used by the people who run it, so it can carry filters, a detailed chart, a data table and the actions they take every day.
The guide asked teams to name their scope in the brief before designing anything. A dashboard trying to serve both audiences tends to end up with too much detail for one and too little for the other.
Across several services
High-level. Links out.
One service
Deeper. More actions.
Choosing charts
Chart questions came up more than any others. Each chart type got its own page with the same parts: what it's for, when to pick something else, the accessibility checks that apply to it, and the charting-library component to start from. Teams could compare two options side by side, and land on the same components other teams were already using.
Line chart
Use it for
Change over time, or comparing a trend against a target.
Pick something else when
There are only a few data points, or more than four lines.
Accessibility
- ✓Label lines directly, not only in a legend
- ✓Vary line style as well as color
Start from
Line chart component
Bar chart
Use it for
Comparing amounts across categories.
Pick something else when
The order of time matters more than the size of each bar.
Accessibility
- ✓Start the axis at zero
- ✓Sort bars unless the order means something
Start from
Bar chart component
Table
Use it for
People need exact values, or need to find one item.
Pick something else when
The pattern matters more than the numbers.
Accessibility
- ✓Right-align numbers
- ✓Keep column headers sortable by keyboard
Start from
Data table component
One place for everything else
The answers teams needed already existed across several design systems, charting libraries and years of research. I didn't rewrite them. The resources page linked to each one with a line on when to use it, showed where the templates and example files lived, and named who to contact for what.
I gathered research from across the company into one collection, and asked teams who had built good dashboards to share their design files so others could start from them. I kept the guide current for four years, through two design-system generations.
- 1. Start with a brief
- 2. What goes first
- 3. Scope
- 4. Choosing charts
- 5. Components
- 6. Accessibility
- 7. Research and examples
- Resources
Resources
Everything this guide points to, in one place. Start here before building something new.
Components
- Data tableSorting, pagination, empty and loading states.
- Charting libraryLine, bar, area and single-number charts.
- Filter barThe shared filters used on list and dashboard pages.
Standards and research
- Accessibility for data visualizationColor, contrast, labels and keyboard access.
- Research collectionPast studies on dashboards, tagged by topic.
- Chart color palettesTested sets for up to eight series.
Files
UX brief template
Shared drive › Dashboards › Templates
Example dashboards
Shared drive › Dashboards › Examples
Chart page template
Shared drive › Dashboards › Templates
Who to contact
UX office hours
Bring a draft or a question. Weekly.
Data visualization working group
Share work and patterns. Monthly.
Charting library team
Bugs and component requests.
Where it landed
Nearly 800 people used the guide, and it became the source of truth for dashboard design across many teams in the org. When the org relaunched its official website, it linked to the guide as a core resource. I became the org's go-to for data visualization, which led to a working group for people doing that work.