A dashboard designed to tell a specific story is very different from one designed for open-ended exploration. Both have their place — but conflating them produces dashboards that do neither well. This guide covers the design principles for each mode.
There are two fundamentally different purposes a dashboard can serve: telling a specific story, or enabling open-ended exploration. Both are legitimate. Both have distinct design requirements. The problem most organisations face is that their dashboards try to do both — and end up doing neither well. This guide covers how to design for each mode and when to use which.
What storytelling dashboards do
A storytelling dashboard communicates a specific insight or analysis to an audience that has a limited time window and a decision to make. The executive briefing. The weekly operations review. The investor metrics summary. The post-campaign performance report.
The viewer of a storytelling dashboard is passive — they arrive, they read, they leave with a conclusion. They do not filter. They do not drill down. They may not even know how to interact with Tableau. The dashboard does the analytical work; the viewer receives the output.
Design principles for storytelling dashboards:
**One insight per screen**: The primary message should be immediately legible before the viewer reads any detail. If you need to look at the dashboard for 10 seconds before understanding the point, the design is working against you. Use a large headline metric, a trend line that makes the direction obvious, and supporting context in smaller elements below.
**Guide the eye with hierarchy**: Size, colour, and position create reading order. The most important element should be largest, in the upper-left, and in the accent colour. Supporting elements are smaller, lower, and in neutral tones. The reader's eye follows the visual hierarchy; design that hierarchy intentionally.
**Answer one question per section**: "What happened to revenue this week?" is one section. "What drove the change?" is the next section. If a section answers two questions simultaneously, split it.
**Minimise interactivity**: Filters and parameter controls on a storytelling dashboard are distractions. The viewer did not come to filter — they came to see the conclusion. If you find yourself adding filters to a storytelling dashboard, ask whether those filters represent questions that belong in a separate exploration dashboard.
**Fixed date ranges, not relative filters**: A storytelling dashboard for an executive review of Q2 performance should show Q2 data. Not "last 90 days" — Q2. Fixed date ranges prevent the dashboard from showing different data depending on when it is opened, which creates confusion in meeting reviews.
**Annotation > interactivity**: If the data had an interesting event (a product launch, a price change, an outage), annotate it directly on the chart. Reference lines, callouts, and custom labels communicate context without requiring the viewer to interact.
What exploration dashboards do
An exploration dashboard is a tool for analysts and power users who want to ask their own questions against the data. They arrive with a hypothesis and leave with either confirmation or a new hypothesis. They filter heavily. They compare segments. They drill from summary to detail.
The viewer of an exploration dashboard is active — they drive the analysis. They know Tableau. They will use every filter, every parameter, every action. They want flexibility, not guidance.
Design principles for exploration dashboards:
**Start at summary, drill to detail**: The exploration begins with an overview — total revenue, trend, top-level segments. Clicking into a specific segment, region, or time period should refine all views on the dashboard. Set actions and filter actions implement this drill-through behavior.
**Consistent filter panel**: Filters should be organised, labelled, and grouped logically. Date range first. Geographic hierarchy next. Product hierarchy next. Status or type filters last. Filters should be visually separated from the charts they control — a left or top filter panel, not filters scattered between charts.
**Show what data supports each view**: For analysts who will use the exploration dashboard for decision-making, they need to understand the data beneath each chart. A tooltip showing the row count behind an aggregate, a reference line showing the historical baseline, a note about data refresh timing — these build analyst trust in the numbers.
**Let users control the dimensions**: Parameter actions and set actions allow analysts to dynamically change what dimension a chart is broken down by. "Break down by region / by segment / by channel" as a parameter selector gives analysts flexibility without requiring separate charts for each breakdown.
**Performance matters more here**: Exploration dashboards are interactive — every filter change triggers a query. An exploration dashboard with 10 filters that takes 30 seconds to load on each filter change will not be used. If the data source is large, use an extract, implement partitioning, and consider row-level aggregation in the data model before the dashboard layer.
The hybrid anti-pattern
The most common dashboard failure mode is the hybrid: a storytelling intent (communicate Q2 performance to leadership) implemented with exploration features (date range slider, 8 segment filters, product hierarchy drill-down). Leadership ignores the filters. Analysts use the storytelling layout (single headline metric, no detail) and find it insufficient for analysis. The dashboard serves neither audience well.
Separate them. Build the storytelling dashboard for the executive audience: fixed Q2 dates, one insight per section, no filters. Build the exploration dashboard for analysts: full date range selector, all segment filters, drill-through actions. They can share the same data source but serve very different needs with very different designs.
When one dashboard has to serve both
Sometimes the budget or timeline does not allow separate dashboards. In that case, design the storytelling layer as the default view, with an "Analyst Mode" button that reveals filters and detailed views. Navigation actions (go to tab) can implement this: Tab 1 is the executive view (locked, annotated, no filters visible), Tab 2 is the analyst view (full filter panel, drill-through). Both tabs use the same data source. The executive sees Tab 1; analysts navigate to Tab 2.
Tableau-specific implementation tips
**Floating layout for storytelling**: Use floating container layout for storytelling dashboards to place elements with precise visual hierarchy control. Use tiled layout for exploration dashboards where filter panels and chart areas need reliable resizing behavior.
**Device previews**: A storytelling dashboard used in leadership meetings is typically viewed on a laptop in full-screen Tableau. A storytelling dashboard embedded in a business review portal may be viewed on a small monitor. Check both. The device designer in Tableau Desktop lets you define separate layouts for desktop, tablet, and phone.
**Guided analytics using Story Points**: Tableau Story is designed exactly for narrative sequencing — a linear flow of dashboards, each advancing the story with annotations and callouts. If your storytelling use case requires a sequence (Context → Problem → Evidence → Recommendation), Tableau Story implements it without custom navigation.
For the visual design foundations, see tableau dashboard design best practices. For the performance considerations for exploration dashboards, see tableau workbook optimization. Our Tableau consulting practice designs both storytelling and exploration dashboards for enterprise clients — book a scoping call to discuss your requirements.
A former Microsoft data architect audits your data foundation, identifies your top priorities, and sends you a written plan. Free. No pitch.
Book a Call →