Information Architecture (IA)
Definition
The structural design of shared information environments — how content is organized, labeled, and connected so that people can find and understand it. Sits at the intersection of library science, UX, and software design.
Often called the blueprint layer of a digital product or knowledge system: invisible when good, painful when bad.
The four IA systems (Rosenfeld & Morville)
| System | What it does | NOTQIN example |
|---|---|---|
| Organization | How content is grouped (taxonomies, classifications) | Folder structure (02_Knowledge Base/Industrial Intelligence/Concepts/), type: frontmatter values |
| Labeling | What things are called | File names, wikilink names, audience-tag values |
| Navigation | How users move through content | Sidebar, breadcrumbs, backlinks panel, portal nav |
| Search | How users find specific things | Obsidian search, Quartz full-text, future RAG over vault |
Why it matters for NOTQIN
The vault’s structure IS information architecture. Decisions like:
- Where to put a new concept note (which folder under
02_Knowledge Base/?) - Whether to use
Customer Anatomy PackorMaster — Customer Anatomyas the canonical link - Whether sector packs go in
01_Project Portfolio/or02_Knowledge Base/ - Audience routing as a parallel IA dimension (per Portal Audience Routing 2026)
…are all IA decisions. Get them wrong and the team won’t find what they need; get them right and the system feels effortless.
Three IA tensions to resolve consciously
- Hierarchy vs network — deep folders (hierarchy) vs flat + wikilinks (network). The vault uses both: shallow folders + dense wikilinking.
- Imposed vs emergent taxonomy — define
type:values up front, OR let them emerge from usage. The vault uses controlledtype:values + free-form tags. - Findability vs discoverability — finding what you’re looking for (search) vs encountering what you didn’t know to look for (browse). Good IA serves both.
Common mistakes
- Letting folder structure drift (every contributor adds their own preferred location)
- Inconsistent labels (
Smartex Auto Tier-2 SprintvsNOTQIN Auto Tier 2 sprint 2026for the same thing) - No controlled vocabulary for
type:or tags → search becomes useless - Over-nested folders (>3 levels deep, nobody navigates)
- Treating IA as a one-time setup vs ongoing maintenance
- Building navigation for the current user; failing to design for the new user
How IA is enforced in this vault
type:frontmatter values are bounded:reference,concept-reference,strategy,playbook,tracker,feedback,log,index, etc. (See NOTQIN KM System — Master Index for the canonical list.)- Hub-entity wikilinking standing rule (every person/company/concept gets
wikilinkform — never plain text) - Folder-level READMEs at each major subfolder explain the IA decision
- Portal Audience Routing 2026 is meta-IA — IA of who sees what
See also
- Knowledge Management Systems — Foundations
- Zettelkasten Method — the network-favoring alternative to hierarchical IA
- NOTQIN KM System — Master Index
- Portal Audience Routing 2026