NoteTaking & Knowledge Management

Tag Taxonomies that Dont Turn into Junk

Create Tag Taxonomies that Don’t Become Junk


🧭 What & Why

What is a tag taxonomy?
A tag taxonomy is a controlled set of labels (and rules for using them) that you apply to notes, documents, and posts so you can find, filter, and connect information fast. It differs from ad-hoc tagging (“folksonomy”) because it’s intentional, curated, and governed—not just whatever comes to mind when you hit “Save.”

Why it matters:

  • Findability & information scent. Good tag names help you predict what’s behind a link or filter; vague or cute tags break that “scent” and cost time.

  • Lower cognitive load. Working memory is limited; fewer, clearer options reduce decision fatigue and errors.

  • Scale without chaos. A small set of reusable tags scales better than hundreds of one-offs; controlled vocabularies address synonyms, homonyms, and spelling drift.


✅ Quick Start: Do This Today

Goal: Clean, reusable tags that help you find things in seconds.

  1. Inventory your current tags. Export or list all tags; sort by frequency. Delete obvious mistakes (<2 uses, typos).

  2. Pick a home framework (choose one): PARA, Johnny.Decimal, or Zettelkasten. You’ll use it to shape categories and tag “families.”

  3. Write 6 Tag Rules (print them):

    • Use singular nouns (habit, recipe, contract).

    • Prefer plain language users would search.

    • One concept per tag; avoid slashes/commas.

    • No project names as tags (make them categories/folders).

    • Min 3-use rule: a tag must appear on ≥3 items within 30 days or it’s culled.

    • Synonym map: choose one canonical tag; redirect others during review.

  4. Create a Starter Set (30–60 tags max). Start with 6–8 “families” (e.g., people, place, time, topic, status, difficulty).

  5. Add an “Incubator” list. New tags go here first; only promote after they hit 3+ uses or add clear retrieval value.

  6. Schedule a 30-minute monthly review (merge, rename, delete, promote).

  7. Template it. Add a “Tags:” line to note/article templates to nudge consistent choices.


🧗 30-60-90 Day Habit Plan

Outcome: A tidy, stable taxonomy with high reuse and low waste.

Days 1–30 (Stabilize):

  • Freeze tag creation outside the Incubator.

  • Apply the 3-use rule to prune.

  • Rename for clarity (e.g., “wellbeing” → “health”).

  • Metric: Total tags < 60, Top 12 cover ≥ 70% of items.

Days 31–60 (Refine):

  • Merge synonyms (e.g., “todo,” “to-do,” “tasks” → “task”).

  • Add attribute tags you truly need (e.g., status: “draft,” “review,” “final”).

  • Introduce hierarchical families if your tool supports it (e.g., topic/habits).

  • Metric: Reuse rate ≥ 80% (8 in 10 items use an existing tag).

Days 61–90 (Scale & Automate):

  • Lock the Core List; gatekeep with monthly reviews.

  • Automate: templates with common tags; bulk-apply via saved searches.

  • Metric: Tag growth ≤ 10%/month with no drop in reuse rate.


🧠 Techniques & Frameworks (and how tags fit)

PARA (Projects, Areas, Resources, Archives)

  • Use: Make PARA your category/folder structure.

  • Tags complement: Add cross-cutting attributes (e.g., time:weekly, status:draft, topic:habits).

  • Why it works: It keeps action close and pushes reference into Resources, avoiding overloaded tag sets.

Johnny.Decimal

  • Use: Give every file/note a stable ID (12.34) under Areas/Categories.

  • Tags complement: Add human-readable facets (client:acme, status:final) so you can filter across IDs.

  • Why it works: The numeric spine prevents sprawl; tags become lightweight filters, not a second filing system.

Zettelkasten

  • Use: Create atomic notes and connect them with links and structure notes.

  • Tags complement: Keep tags minimal (e.g., concepts or statuses) and let links do most of the connecting.

  • Why it works: The web of connections is primary; tags are supporting actors.

Information Scent & Naming

  • Write tags the way a searcher thinks (“budget,” not “pecuniary prudence”).

  • Prefer short, concrete words; avoid jargon unless your audience expects it.

Controlled Vocabulary vs. Folksonomy (in PKM)

  • Controlled vocabularies reduce ambiguity (synonyms, homonyms) and improve recall.

  • Folksonomies surface emergent language but drift without governance.

  • Best of both: controlled core + curated incubator for new terms.


🛠️ Tag Decision Matrix (use this every time)

Question If Yes If No
Will I use this tag on ≥3 items this month? Add to Incubator; promote after 3 uses. Don’t create it. Use an existing tag.
Does this tag change how I’ll find things later? Keep/merge synonyms into one canonical tag. Remove; it’s decorative.
Is the tag a status or attribute (draft, client, level)? OK—these cut across folders. Consider folder/category instead.
Is the wording obvious to future-you? Keep. Rename for clarity (short, singular noun).

👥 Audience Variations

Students:

  • Use course code as a folder/category; tag with topic (e.g., topic:lab-reports), exam:midterm, status:draft/final.

  • Weekly tag review during study planning.

Professionals:

  • Use client and status tags; keep topic tags lean.

  • Add decision: tags for meeting notes (decision:approved/rejected).

Researchers/Grad Students:

  • Use method (method:RCT, method:qual) and population tags; keep theory names canonical (theory:self-determination).

  • Maintain a synonym map (e.g., “SDT” → “self-determination”).

Teams:

  • Publish a one-page Tag Policy; appoint a “taxonomy steward.”

  • Run a monthly merge meeting; log changes in a changelog.

Seniors/Low-tech:

  • Keep ≤20 tags; focus on people, places, bills, health, family.

  • Use voice or checklists to apply standard tags.


⚠️ Mistakes & Myths to Avoid

  • Myth: More tags = better search. Reality: more choices increase cognitive load and slow you down.

  • Mistake: Clever names. If it’s not obvious, it’s invisible. Maintain information scent.

  • Mistake: Project names as tags. Use folders/categories for projects; tags should cut across projects.

  • Mistake: Never deleting tags. Prune aggressively; archive old tags in a changelog.

  • Mistake: Mixing plural/singular or tense. Pick one convention (prefer singular nouns).


💬 Real-Life Examples & Scripts

1) Monthly Tag Review Agenda (30 minutes)

  • 5 min: Sort by frequency; delete or demote tags with <3 uses.

  • 10 min: Merge synonyms (choose the clearest; update past items).

  • 10 min: Promote Incubator tags that hit criteria; add to Core List.

  • 5 min: Update the Tag Policy and changelog.

2) Synonym Decision Script

“Between wellbeing, well-being, and health, which would future-me type to search? Choose one canonical: health. Merge others and note the redirect in the policy.”

3) Naming Script

“Could a newcomer guess what this tag means without context? If not, rename to a shorter, plainer term.”

4) PARA Mapping Example

  • Projects: “2025 Habit Course” (folder)

  • Areas: “Editorial,” “Finance” (folders)

  • Resources: “Research—Habits” (folder)

  • Archives: finished projects (folder)

  • Tags used across all: topic:habits, status:draft/final, audience:students.

5) Johnny.Decimal Mapping Example

  • 11.02 Editorial Planning (category) → Notes carry tags like month:oct, status:ready.

6) Zettelkasten Example

  • Keep tags minimal (concept:identity-based-habits), rely on links and structure notes to weave ideas.


🧰 Tools, Apps & Resources (quick pros/cons)

  • Obsidian / Logseq / Tana / Roam – rapid linking and tags; easy bulk edits; great for Zettelkasten-style networks.

  • Notion / Evernote / OneNote – simple tagging, good templates; watch for silent drift in team spaces.

  • DEVONthink – powerful AI-assisted filing; tags plus smart groups.

  • WordPress – use categories as “PARA-like” containers; keep tags curated (avoid one-off SEO spam).

  • Governance helpers (any platform): saved searches for low-use tags, bulk-rename/merge tools, templates that include a “Tags:” line by default.

Design tip: Regardless of platform, define Core Tags, a Tag Policy (1 page), and a Review Cadence (monthly). The combo is more important than the tool.


📚 Key Takeaways

  • Constrain choices: a small, controlled list outperforms sprawling folksonomies for everyday retrieval.

  • Choose a backbone framework (PARA, Johnny.Decimal, Zettelkasten) and let tags be lightweight, reusable attributes.

  • Protect information scent with plain, predictable names.

  • Run monthly reviews to merge, promote, or delete; keep an Incubator to test new tags.

  • Track simple health metrics: total tags, reuse rate, and how much your top 12 cover.


❓ FAQs

1) How many tags should I have?
Start with 30–60 and let real usage guide you. If your top 12 tags don’t cover most items, you’re likely over-tagging.

2) What’s the difference between categories and tags?
Categories (or folders) provide primary structure; tags are cross-cutting attributes that help filtering and search. Use both; don’t duplicate.

3) Should I use hierarchical tags (e.g., topic/habits)?
If your tool supports them and your list is large, yes—families reduce ambiguity and speed scanning.

4) Do I need plural or singular names?
Pick singular nouns for consistency and easier autocomplete (e.g., habit, not habits).

5) How do I stop new tags from exploding?
Adopt an Incubator, a 3-use rule, and a monthly review. Promote only what proves useful.

6) Are folksonomies bad?
They’re great for discovering emergent language but need governance; a controlled core + incubator balances both.

7) How do naming choices affect findability?
Clear, predictable names preserve information scent, making links/filters easier to follow.

8) What metrics should I watch?
Total tag count, reuse rate (items using existing tags), and “top 12 coverage” (% of items using your dozen most common tags).

9) Where should “project names” live—category or tag?
Make projects categories/folders (they begin and end). Use tags for timeless attributes.

10) How do frameworks help?
PARA/Johnny.Decimal/Zettelkasten stabilize structure so tags don’t carry the whole load, keeping your taxonomy lean.


📚 References