Usability · A practitioner's guide
What is usability, and why does it matter?
Usability is the quality that makes a product easy, efficient, and satisfying to use. It is not one thing — it is a composite of learnability, efficiency, memorability, error tolerance, and satisfaction. Get it right and people accomplish their goals. Get it wrong and they leave, or make costly mistakes.
01 — The components
The five qualities under one word.
Usability is a composite. Lose one component and the whole experience tilts — even when every other dimension is strong.
Learnability
First-use task success
How quickly can a new user accomplish basic tasks on first encounter?
Efficiency
Speed once learned
Once learned, how fast can experienced users reach their goals?
Memorability
Returning after absence
When users return after time away, can they re-establish proficiency quickly?
Error tolerance
Recovery cost
How severe are mistakes, and how easily can users recover from them?
Satisfaction
Subjective experience
Is the experience pleasant and fulfilling, not just functional?
UX is not UI
Surface vs experience
UI is the surface; UX is the entire experience — the moment of confusion, the moment of success, the feeling after the session ends.
02 — Recruiting
The legend of 'the general public.'
Research aimed at 'anyone' produces findings that apply to no one. Every product has a real user — a person with domain knowledge, a mental model shaped by prior tools, and specific goals.
- 01
Define the user profile first — Role, domain knowledge, frequency of use, technical comfort — be specific before opening any recruiting tool.
- 02
Write a tight screener — 6–8 questions max. Avoid leading language. Screen on behaviours, not demographics alone.
- 03
Use multiple channels — Customer success teams, LinkedIn, existing panels, UX agencies. Do not rely on a single source.
- 04
Plan for 5–8 participants per segment — Diminishing returns kick in fast; beyond 8 you are confirming, not discovering.
- 05
Compensate fairly — Poor incentives bias your sample toward people who need the money, not your target user.
Specificity in recruiting is the difference between research that changes product decisions and research that validates whatever stakeholders already believed.
03 — Methods
Qualitative vs quantitative · attitudinal vs behavioural.
No single method answers every question. The choice depends on where you are in the design process and what you need to know.
Qualitative (Why)
Moderated testing, contextual inquiry, cognitive walkthroughs. Use when you need to understand why something is happening.
Quantitative (How many)
Unmoderated testing, surveys, analytics, A/B testing. Use when you need to measure scale and frequency.
Attitudinal (What they say)
Surveys, NPS, interviews, focus groups. Captures stated preference, perception, satisfaction.
Behavioural (What they do)
Moderated and unmoderated testing, analytics, eye tracking. Prefer this when you can — what people say and what they do are often different.
Method reference
- 01
Card sorting — Reveals mental models for navigation. Run early.
- 02
Tree testing — Validates IA without visual design interference. Run after card sorting.
- 03
Diary studies — Captures longitudinal experience over days or weeks. Costly but uniquely revealing.
- 04
Heuristic evaluation — Fast expert review. Surfaces obvious issues before participant studies.
- 05
First-click testing — Validates whether users find the right starting point for a task.
Mix generative and evaluative methods. Generative tells you what to build; evaluative tells you whether you built it well.
04 — Research velocity
A 5–7 day research sprint.
A 3-day study with clear output beats a 3-week study with a 60-slide deck nobody finishes reading.
Day 1
Research Brief
Define the research question. Align stakeholders. Not the solution — the question.
Days 2–3
Study plan + screener
Method, participant criteria, session guide, schedule, roles. Written and approved.
Day 3
Pilot session
Run once with a colleague. Fix broken flows and timing before real participants.
Days 4–5
Sessions (5–8 participants)
Run sessions. Synthesise same-day — cluster observations within 24 hours.
Day 6
Synthesis
Themes, severity, prioritisation. Do not batch to end of study.
Day 7
Two-format readout
1-page summary for decisions. Full report for the record. Lead with the summary.
05 — Guidelines
Nielsen's 10 heuristics.
Heuristics are a diagnostic vocabulary, not a design spec. Severity is context-dependent — a violation affecting 80% of users daily dwarfs one affecting 0.1% weekly.
- 01
Visibility of system status — Always keep users informed about what is happening.
- 02
Match the real world — Use words and concepts the user knows.
- 03
User control & freedom — Support undo, redo, and easy exits.
- 04
Consistency & standards — Follow platform conventions.
- 05
Error prevention — Design to prevent problems first.
- 06
Recognition over recall — Make options visible; do not make users memorise.
- 07
Flexibility & efficiency — Accelerators for experts; do not punish novices.
- 08
Aesthetic & minimalist design — Every extra element competes for attention.
- 09
Recognise & recover from errors — Plain language, precise, constructive.
- 10
Help & documentation — Easy to find, action-oriented.
Severity scale for heuristic violations
| Severity | Definition | Action | Example |
|---|---|---|---|
| Critical (4) | Prevents task completion. High frequency. | Fix before release | Form submit disabled with no feedback. |
| Serious (3) | Causes significant confusion or delay. | Fix this sprint | Error message says "Error 403" only. |
| Minor (2) | Irritating but user can work around it. | Schedule soon | Inconsistent button label between two screens. |
| Cosmetic (1) | Noticeable but almost no impact on task. | Fix if time allows | Slight misalignment on a rarely used modal. |
| Not an issue (0) | Disputed or not a real problem. | Document, move on | Reviewer preference vs. established standard. |
06 — Content strategy
Content is the product.
Words, labels, instructions, and error messages are interface elements — not filler. Content strategy in UX means designing the information architecture, hierarchy, and language from the first wireframe.
- 01
Lead with the user's goal — Labels, headings, and CTAs should answer 'what does the user want to do?' not 'what does this section contain?'
- 02
Write at the level of your audience — Plain language reduces cognitive load for everyone. Jargon is acceptable when the audience genuinely owns it.
- 03
Error messages are content — 'Something went wrong' is not a message. Tell users what happened, what it means, and how to fix it.
- 04
Audit before redesign — Inventory what exists, what's redundant, what's missing before redesigning anything.
- 05
Test your labels — Card sorting and first-click tests validate whether labels are interpreted as intended.
"We'll fix the copy later" is how products ship with confirmation dialogs saying "Are you sure?" and errors saying "Error 403." Make content a first-class design decision.
07 — Visual design
Hierarchy is the usability tool.
Visual design is not decoration — it is communication. Every spacing decision, contrast ratio, and typographic weight either helps users understand a hierarchy or makes it harder.
- 01
Hierarchy before aesthetics — Define primary, secondary, and tertiary actions before choosing colours. The most important action should be most visually prominent.
- 02
WCAG AA as a floor — 4.5:1 contrast ratio for text is the minimum. Aim for 7:1 in data-dense contexts.
- 03
Don't communicate with colour alone — 8% of males have some form of colour blindness. Use shape, label, and pattern alongside colour.
- 04
Whitespace is information architecture — Spacing groups related elements and separates unrelated ones. Insufficient whitespace makes everything feel equally important.
- 05
Typography is UI — Font size, weight, and line-height determine whether users can scan efficiently. 16px body, 1.5 line-height, limited weight range.
A usability test revealing users ignoring your primary CTA is often a visual design problem. Hierarchy, contrast, and size communicate intent before the user reads a single word.
08 — Research ops
Managing UX research like a product.
UX research does not manage itself. Without a clear plan, studies run over schedule, findings get buried in decks nobody reads, and stakeholders lose confidence in the process.
- 01
Research brief (day 1) — Document the research question, not the solution hypothesis. Align stakeholders before recruiting starts.
- 02
Study plan (week 1) — Method, participant criteria, screener, session guide, schedule, roles. Written, reviewed, approved.
- 03
Pilot session — Run one session with a colleague first. Catch broken prototypes and timing issues.
- 04
Same-day synthesis — Cluster observations within 24 hours. Do not batch synthesis to the end of a two-week study.
- 05
Two-format readout — A one-page summary for quick decisions. A detailed report for the record. Lead with the one-pager.
- 06
Track decisions, not just findings — What changed because of this research? Document it. This is how research builds credibility over time.
Research velocity matters. Treat research like a product: scope it, schedule it, ship findings on time.