BlogData Analytics SeriesChapter 22
SERIES · CHAPTER 22Professional Skills

Dashboard Design Best Practices for Data Analysts

Layout hierarchy, KPI tile design, colour rules, chart selection, removing clutter, mobile-first design, and before/after redesigns — with Power BI examples and Indian business dashboard patterns. Build dashboards that get used every day, not opened once and forgotten.

DATA ANALYTICS SERIES:← Ch 21: Data StorytellingCh 22: Dashboard Design ←Ch 23: BI Strategy →

5 Principles of Effective Dashboard Design

1
Hierarchy — most important information top-left
The human eye scans in an F-pattern (top, then left). KPI summary tiles go top-left. Detail charts go below. Drill-down tables go at the bottom or in a secondary tab. Every element on the page should earn its position by importance — not by the order you built it.
2
Context — every number needs a comparison
A KPI tile showing "GMV: ₹4.2Cr" is useless without context. Is that good or bad? Add: vs target, vs last week, vs last year, or a sparkline showing trend. A number without context is a fact. A number with context is an insight.
3
Colour communicates — use it intentionally
Red = below target or declining. Green = above target or growing. Use one accent colour for the most important metric. Grey out secondary elements. Never use colour just to make the dashboard look busy. Every colour choice should have a reason.
4
Simplicity — remove everything that does not earn its place
Every gridline, border, 3D effect, logo, or decorative element that does not convey data is cognitive noise. Remove it. A dashboard that looks "bare" to you likely looks exactly right to a busy stakeholder who has 30 seconds to understand performance.
5
Consistency — same colours, fonts, and patterns throughout
If Revenue is always blue and Returns are always red across every chart on the dashboard, viewers learn the system once and never have to re-check the legend. Inconsistency forces re-reading every chart from scratch.

Dashboard Layout Blueprint

STANDARD EXECUTIVE DASHBOARD LAYOUT (e.g. Power BI, Looker)
GMV ₹4.2Cr ▲12% WoW
Orders 34,821 ▲8% WoW
Return Rate 18.3% ▼0.4pp
New Customers 12,440 ▲5% WoW
Row 1: KPI Tiles — 4–6 tiles, top of page, with comparison values and colour coding
Primary Trend Chart
(Revenue or Orders over time — line chart)
Category/Channel Split
(Bar or Donut chart)
Row 2: Primary charts — trend + composition. 2/3 + 1/3 split.
Geo Map (orders by state)
Funnel (conversion steps)
Table (top 10 products)
Row 3: Supporting detail — geography, funnel, ranking table. Below the fold.

KPI Tile Anatomy — What Every Tile Needs

🏷️
Metric name
Short, unambiguous. "GMV" not "Gross Merchandise Value (Including Cancelled Orders Before Return)". If you need a tooltip, the name is too complex.
🔢
Current value
Large font. Unit explicit (₹ Cr, %, days). Round appropriately — "₹4.2Cr" not "₹4,21,34,523.00".
📊
Comparison value
vs target, vs last period, or vs last year — one comparison only. "▲12% vs last week." Choose the comparison that drives the business decision.
↕️
Direction indicator
Up/down arrow. Colour: green if improvement, red if decline. Note: for return rate, an UP arrow should be red (higher returns = worse).
〰️
Sparkline (optional)
A tiny 30-day trend line below the number. Shows whether today's value is typical, trending, or anomalous. Worth adding when the single number can be misleading without context.
🕐
Last refresh time
Bottom of the dashboard: "Last updated: 07 Sep 2026, 06:30 IST." Without this, users cannot trust whether the data is stale.

Dashboard Anti-Patterns — What to Stop Doing

Anti-PatternWhy It FailsFix
Putting 20+ metrics on one pageCognitive overload — nothing stands out, nothing gets actionedMaximum 6 KPI tiles; remaining metrics go in sub-pages or appendix
Pie charts with 8 slicesImpossible to compare slice sizes; legend mapping is exhaustingUse a horizontal bar chart; group small slices into "Other"
3D charts of any kind3D distorts proportions and adds no informationDelete all 3D charts immediately — no exceptions
Traffic-light KPIs with no thresholds explainedUsers do not know what "amber" means for each metricAdd tooltip or footnote: "Red = >10% below target; Amber = 5–10% below"
No date filter visibleViewer cannot tell if they are looking at today, this week, or this yearAlways show the active date range prominently near the top
Colour every bar a different colourForces constant legend-checking; colours should group, not labelUse one colour for all bars; highlight the one bar that matters
Showing percentages without absolute numbers"Return rate up 2pp" — but was it 1,000 returns or 100,000?Show both: "18.3% (31,840 orders)"
No mobile testingDesktop dashboard looks fine; on phone, 3-column KPI row becomes a horizontal-scroll nightmareBuild a dedicated mobile layout in Power BI; test on actual Android device
Continue the Series
← Ch 21: Data StorytellingCh 23: BI Strategy →

Frequently Asked Questions

How many KPIs should a dashboard show?

A well-designed executive dashboard shows 4–6 KPI tiles at the top. This is not an arbitrary rule — it comes from cognitive load research showing that humans can hold approximately 4 items in working memory simultaneously. More than 6 KPIs on a single view means the user cannot monitor all of them at once; important signals get missed. If you have 15 metrics to track, create multiple dashboards: an executive summary dashboard (4–6 KPIs, daily use), an operational detail dashboard (8–12 metrics, weekly use), and deep-dive dashboards per business function. Each audience gets the dashboard calibrated for their decision frequency and detail level. In Indian corporate practice, the common mistake is to put every metric available into one dashboard to demonstrate thoroughness — this optimises for analyst effort, not for stakeholder comprehension. A dashboard that shows less but shows it clearly will be used every day. A dashboard that shows everything will be opened once and never revisited.

What colours should a data analyst use in dashboards?

Colour in dashboards should serve a specific purpose — not decoration. Three practical colour rules: (1) Use one primary brand colour for positive or neutral values; use red only for negative performance or alerts. Never use red and green arbitrarily — approximately 8% of men have red-green colour blindness. (2) Limit your categorical palette to 5–6 colours maximum. More than 6 categories on a single chart requires a legend and forces the reader to match colours to labels — this breaks visual flow. If you have 10 categories, group the smaller ones into "Other." (3) Use sequential palettes (light to dark of one hue) for showing magnitude (revenue by state — darker = higher). Use diverging palettes (two contrasting hues with neutral centre) for showing deviation from a baseline (% above/below target). In Power BI and Looker, use the conditional formatting feature to colour KPI tiles: green if above target, amber if within 10% below, red if more than 10% below. This gives instant visual triage without the reader needing to read numbers.

Should analytics dashboards be designed for mobile in India?

Yes — especially for operational dashboards viewed by field teams, delivery managers, store managers, and sales representatives in India. Mobile penetration in Indian enterprises is high, and many operational roles access dashboards exclusively on phones. For mobile-first dashboards: (1) Stack KPI tiles vertically (one column) rather than a 3-column grid that requires horizontal scrolling on phones. (2) Avoid data tables with many columns — they are unreadable on mobile. Use summary tiles or simple bar charts instead. (3) Use larger font sizes — minimum 14px for body text in charts, 20px for KPI values. (4) Limit interactive filters to 1–2 slicers per page on mobile views. (5) Test your Power BI or Looker dashboard on an actual Android phone before publishing — what looks fine at 1440px desktop width often breaks on a 360px mobile screen. Power BI has a dedicated "mobile layout" editor where you can arrange visuals specifically for phone viewing without affecting the desktop layout.

What is the difference between a dashboard and a report in data analytics?

A dashboard is a live, interactive view of current performance — designed for frequent monitoring (daily or real-time). It answers: "How are we doing right now?" It should load quickly, update automatically, and communicate status at a glance without requiring any analysis from the viewer. A report is a structured document (static or refreshed periodically) that analyses a specific question in depth — with context, methodology, findings, and recommendations. It answers: "What happened, why, and what should we do?" Reports are shared at a specific point in time (weekly business review, quarterly analysis) and are designed to be read, not monitored. Confusion between the two leads to bad design: overly detailed dashboards that require reading rather than glancing, or one-page reports that lack the context needed to understand the data. In Indian company workflows: the BI analyst typically owns dashboards (Power BI, Looker, Tableau); the data analyst produces reports (slide decks, notebooks, SQL readouts). Both require storytelling; dashboards require design thinking that reports do not.

EVIKA ACADEMY · NOIDA SECTOR 51

Build Dashboards That Drive Decisions

Our Power BI curriculum covers dashboard design, DAX measures, and live project work building real operational dashboards — with design feedback and stakeholder presentation practice.

Book Free Demo Class →
🎓 Free Demo Class — Online & Offline · Noida Sector 51