Visualization and UX

This theme evaluates the report's ability to make things understood quickly and without error: choice of visuals, information hierarchy, navigation (bookmarks, tooltips, drill-through pages)…

Guide≈ 5 min readCheckedReading

In short

This theme evaluates the report's ability to make things understood quickly and without error: choice of visuals, information hierarchy, navigation (bookmarks, tooltips, drill-through pages), consistency (JSON theme, template), cognitive load, usability testing. We assess substance (principles, decisions, evidence), not visual polish. Accessibility and multilingual support are covered separately: ACC.

On this page

1. Why this theme
2. What this theme measures
3. The 5 levels
4. Milestones by level
5. Detailed reading by level
6. Partial achievement and rating
7. Moving to the next level
8. See also

Level: 🟢→🔴 · Theme code: VIZ

1. Why this theme#

An accurate number presented poorly is misread; a misleading chart (truncated axis, 12-slice pie chart, dual axis) leads to bad decisions. A 30-page report “showing everything” is only consulted by its author. Cost of ignoring this: low adoption, data extraction requests, multiplication of parallel reports (GOV).

2. What this theme measures#

Axis: from an accumulation of default visuals to a designed, tested, and consistent experience.
Key indicators: time to answer 3 standard questions (usability test), number of clicks to reach information, page usage rate, template compliance.
Units: seconds, clicks, % of compliant pages.

3. The 5 levels#

Level	Short name	Promise	Typical indicator (orders of magnitude)
1	Default	Numbers are displayed; the reader has to guess where to look.	Default visuals, 15+ visuals per page
2	Shared template	All team reports look alike; the reader finds familiar cues.	Shared JSON theme; 1 template
3	Readable journey	A new user answers their 3 main questions in under one minute.	Usability test: < 60 s for 3 questions
4	Design system	The report is consistent, documented, and reusable; a new screen is built in hours, not days.	Component library; design review at each delivery
5	Measured adaptive	The experience adjusts to roles and devices, and design decisions are based on measured usage.	Usage telemetry used; mobile validated

4. Milestones by level#

ID	Milestone	Verifiable evidence
BI-VIZ-1.1	Each page has a clear title and each visual has a title	Page-by-page review
BI-VIZ-1.2	Numbers have a unit and format (thousands, %, currency)	Formats set on measures
BI-VIZ-1.3	No visual is obviously misleading (truncated axes without mention)	Critical review
BI-VIZ-2.1	A shared JSON theme (colors, fonts, sizes)	Theme file in repository
BI-VIZ-2.2	Page template (header, filters, content area)	`.pbit` model / standard page
BI-VIZ-2.3	Visual types chosen based on the question (comparison, trend, distribution, relationship)	Internal selection table
BI-VIZ-2.4	No more than ~10 to 15 visuals per page	Review + PRF
BI-VIZ-3.1	Visual hierarchy: key indicator at top left, detail below	Mockup review
BI-VIZ-3.2	Consistent navigation (page menu, back, bookmarks)	Journey tested
BI-VIZ-3.3	Mockups validated with a user before development	Mockup + report
BI-VIZ-3.4	Reserved semantic colors (red/green for variances, never decorative)	Color guidelines
BI-VIZ-3.5	Page tooltips and drill-through pages for detail without overload	“Drill-through” pages in use
BI-VIZ-4.1	Documented component library (KPI cards, headers, matrix templates)	Component repository
BI-VIZ-4.2	Systematic design review at each delivery	Completed review checklist
BI-VIZ-4.3	Custom visuals limited and approved (security, performance, support)	Visual whitelist
BI-VIZ-4.4	Timed usability test on 3 questions per major report	Test report
BI-VIZ-5.1	Usage telemetry (page views, filters, drop-offs) used to make decisions	Quarterly usage review
BI-VIZ-5.2	Mobile variant designed and tested	Mobile layout verified
BI-VIZ-5.3	Journeys adapted to roles (executive / manager / field)	Pages or reports by role
BI-VIZ-5.4	Systematic removal of little-used pages	Removal log

5. Detailed reading by level#

🟢 Level 1 — Default#

Signs: default visuals, technical titles (“Sum of Amount”), random colors.
Benefit: speed. Does not guarantee: understanding.
Difficulty: 1/5 · Effort: 1/5.

🟢 Level 2 — Shared template#

Signs: shared JSON theme and template, consistent number formats, visual types chosen logically (bars for comparison, lines for trends).
Benefit: consistency and production speed.
Limit: consistency does not imply relevance.
Difficulty: 2/5 · Effort: 2/5.

🟡 Level 3 — Readable journey#

Signs: design starts from user questions, not available data. One page = one message. Details are in tooltips or drill-through, not overload.
Benefit: adoption; fewer extraction requests.

🧪 Example

SFA common thread: the “Field coverage” page answers “are my visits being completed?” (completion rate vs target, by week); customer drill-through is available via drill-through, not shown by default.

Limit: validated by few people.
Difficulty: 3/5 · Effort: 3/5.

⚠️ Pitfall

Symptom: users ask for “an Excel export of the table.” Cause: the matrix does not answer the real question, or it is too dense. Remedy: investigate how the export is used, and make drill-through natural rather than banning export.

🟡 Level 4 — Design system#

Signs: reusable components, tightly controlled custom visuals, timed usability testing, design review.
Benefit: consistency at scale, easier training.
Limit: system maintenance cost.
Difficulty: 4/5 · Effort: 4/5.

🔴 Level 5 — Measured adaptive#

Signs: telemetry (service activity logs) feeds design decisions; mobile and role-based versions; regular cleanup of dead pages.
Benefit: the report evolves with real usage.
Limit: privacy compliance in telemetry (RLS, PD — security).
Difficulty: 5/5 · Effort: 4/5.

✅ Good practice

Ask “what decision does this page enable?”. A page with no associated decision is a candidate for removal.

💡 Expert tip

To keep a theme under control, version the JSON theme in a repository and apply it through the PBIP project (IND); colors “hard-coded” in a visual are technical debt to track down.

🤖 AI

Copilot/assistants can suggest a layout and titles; they reproduce clichés (pie chart, gauge) and do not test anything with a real user.

6. Partial achievement and rating#

Common rule. “A partial 5”: telemetry reviewed but not yet used to remove pages (BI-VIZ-5.1 = 0.5; 5.4 = 0), mobile designed for only one report (5.2 = 0.5), roles half covered (5.3 = 0.5).

7. Moving to the next level#

Possible means (optional):

Transition	Possible levers
1 → 2	JSON theme; template; formatting rules
2 → 3	Mockup workshops; 3-question test; drill-through
3 → 4	Component library; review checklist; visual whitelist
4 → 5	Activity logs used; mobile; role-based journeys

Pitfalls:

False positive: a nice-looking report never tested with a user.
Goodhart: “< 10 visuals per page” leads to multiplying pages; judge the journey, not the count.
A design system that constrains too much: allow justified exceptions.

8. See also#

Related themes: ACC · PRF · GOV · DAX
Other matrices: PD — Adoption and Change Management · PI — Stakeholders and Communication (PAR)
Content: Cross-Cutting Power BI Best Practices · Power BI Reporting Project Checklists · OneLake + SaaS Mobile Apps: Overview

✻ Written by Claude Sonnet 5.5