Designs information architecture by generating sitemaps, navigation hierarchies, and content taxonomies, running card sort and tree test analyses, and auditing navigation labels against findability criteria. Use when structuring a new site/app's content, when users report "I can't find anything," when running a card sort or tree test, or when auditing navigation labels and taxonomy for an existing product.
Resources
1Install
npx skillscat add uw-ssec/savvy/information-architecture Install via the SkillsCat registry.
SKILL.md
Information Architecture
Workflow
- Inventory content — list every page/feature/content-type in a spreadsheet (cols: URL, title, type, owner, last-updated, traffic).
- Identify user goals — top 10 tasks users come to do.
- Run open card sort (15-20 participants via Optimal Workshop / Maze).
- Draft sitemap — depth ≤ 4 levels, ≤ 7 siblings per node.
- Tree test — 10+ tasks, 30+ participants.
- Pass: ≥ 70% direct success (target found within 3 clicks, no backtrack) per task.
- Fail: restructure the failing branch and re-test before shipping.
- Lock labels — first-click test the top 10 labels (≥ 80% clicking expected option).
- Wire to navigation patterns — see navigation-guide.md.
Deliverables (formats in references)
- Sitemap — indented tree, depth ≤ 4, siblings ≤ 7. Example: sitemap-patterns.md.
- Card sort analysis — clusters with agreement %, disputed cards with action, participant vocabulary. Template: card-sorting-methods.md.
- Navigation spec — pattern + primary/secondary/wayfinding + rationale. Format: navigation-guide.md.
Labeling audit checklist
For each navigation label, verify:
- Unambiguous: a stranger can predict what's behind it
- Consistent: same concept = same word across product
- User language: matches card-sort vocabulary, not internal jargon
- Mutually exclusive: no overlap with sibling labels ("Shoes" vs "Footwear")
- Scannable: meaningful word first ("Account Settings" not "Settings for Your Account")
- First-click test success ≥ 80%
Navigation pattern picker
| Pattern | When |
|---|---|
| Hierarchical tree | Large content sites, docs, enterprise — depth ≤ 4 |
| Flat | 4-7 equal-weight sections; mobile bottom-tab |
| Hub-and-spoke | Independent task flows; rare cross-section nav |
| Sequential | Onboarding, checkout, wizard — always show progress + back |
| Faceted | E-commerce, search-heavy libraries with multi-attribute items |
Details: navigation-guide.md.
Validation checkpoints
- After card sort: clusters with < 60% agreement → re-run with refined card set.
- After tree test: any task < 70% success → restructure failing branch, re-test.
- After labeling audit: any label with < 80% first-click success → rename.
- After launch: monitor search queries — high volume of "where is X" = IA failure.
Next Steps
wireframing (the wireframing skill) · user-research (the user-research skill) · ux-writing (the ux-writing skill)