Skip to content
Home » Blog » Blog » Personal Knowledge Management » How to Tag Your Notes So You Can Find Them

How to Tag Your Notes So You Can Find Them

How to tag your notes

Article summary

Most tagging systems fail because they’re built for the moment of saving, not the moment of finding. A few cognitive principles and practical rules can fix that.

  • Tag for your future search query, not for the note’s topic. Object tags (what the note is about) outperform topic tags (what domain it relates to) as your archive grows past a few hundred notes.
  • Use 3-5 tags per note across distinct dimensions. One for note type, one for status, one or two for specific concepts. More creates noise and fewer creates orphans.
  • Schedule monthly tag maintenance. Merge synonyms, retire dead tags, and keep a single master list with short definitions for every active tag.

The tagging system that works

You know it’s somewhere in your notes. You can almost picture the page. But the search bar stares back at you, empty-handed. The note exists. Your ability to retrieve it is lost.

This is the central frustration of any personal knowledge system, and it comes down to one thing:

“How you tagged (or didn’t tag) that note when you captured it.”

Tagging feels like a minor clerical task. It’s not. It’s the bridge between the information you collect and the arguments you build. Get it wrong, and your growing archive becomes a graveyard of ideas you can’t reach.

The good news is that you don’t need a complicated system. You need a few clear principles grounded in how memory and retrieval work, and the discipline to apply them consistently.

Why most tagging fails: The 2 cognitive traps

The reason tagging breaks down isn’t laziness or bad software. It’s a mismatch between two versions of yourself: the you who saves a note and the you who later needs it. These two people think differently about the same information.

Cognitive scientists have a name for this problem. Tulving and Thomson’s encoding specificity principle, published in Psychological Review in 1973, showed that memory retrieval works best when the cues at recall match the cues present during encoding. A tag only helps if it overlaps with how you’ll be thinking about that information when you later want to find it. If you tag a note about insulin resistance with #health because that’s the broad domain you were browsing, you’ve created a cue that won’t fire when you’re drafting a paragraph about metabolic responses to carbohydrates six months later.

The second trap is what Watkins (1975) called the cue overload principle. A retrieval cue loses power the more memories it’s linked to. Tag 400 notes with #research and that tag becomes functionally useless. It points everywhere, which means it points nowhere. Sascha Fast from zettelkasten.de uses a memorable analogy for this:

“Searching a topic tag in your archive is like firing a shotgun into the woods and hoping dinner appears on the table. What you want is a sniper rifle.”

These two principles together explain why the instinct to tag broadly (“I’ll just tag this #psychology”) produces systems that feel organized at first and collapse within months.

Object tags vs. topic tags

The single most useful tagging insight I’ve encountered comes from the Zettelkasten community, specifically from Sascha Fast’s writing on zettelkasten.de.

“Most people tag for topics when they should tag for objects.”

A topic tag marks everything relevant to a subject. A note on insulin sensitivity gets tagged #diet because it’s relevant to diet. A note on sleep architecture gets tagged #health because it relates to health. This feels logical. But when you search #diet later, you get a cloud of tangentially related notes: glycemic index, meal timing, paleo anthropology, gut bacteria. You wanted the insulin note. You got the entire warehouse.

An object tag marks what the note is specifically about. The insulin note gets tagged #insulin-sensitivity. The sleep note gets #sleep-stages. Now your tags are precise. They return exactly what they promise.

This mirrors how semantic memory works. The concept “penguin” activates “bird,” “Antarctica,” and “black-and-white feathers” in your brain. It doesn’t activate “nature documentary” or “David Attenborough.” The strongest retrieval paths connect to close conceptual neighbors, not distant thematic relatives. Your tags should follow the same logic.

For research writing, object tags are particularly powerful. When you’re assembling an argument about, say, the relationship between sleep deprivation and decision-making, you want tags like #sleep-deprivation, #decision-fatigue and #prefrontal-cortex-function. You don’t want to wade through everything you’ve ever saved about sleep or about decisions.

Practical tagging for research notes

The systems that hold up over time share a common structure. They tag along multiple independent dimensions rather than trying to capture everything in one label. Think of it as giving each note a few coordinates on a map instead of dropping it into a single bin.

Dimension 1: Note type

What kind of note is this? Borrowed from Richard Carter’s tagsonomy for Zettelkasten, type tags classify the form of the note, not its content. This avoids the “is this psychology or neuroscience?” problem entirely because you’re not tagging topics at all.

A minimal set:

  • #fleeting (raw thought, unprocessed)
  • #source (highlight, quote, or paraphrase from a specific text)
  • #idea (your own developed thinking)
  • #definition (a concept you’ve clarified in your own words)
  • #index (a meta-note linking to other notes on a cluster)
  • #question (open problems you haven’t resolved)

Dimension 2: Workflow status

Where is this note in your process? Status tags prevent valuable material from going stale. They also let you pull up everything that needs attention without reading every note.

The basics:

  • #inbox (captured, not yet processed)
  • #developing (partially worked through)
  • #evergreen (fully developed, standalone, written in your own words)
  • #archived (no longer active)
  • For article writing specifically, you might add #supports-thesis, #challenges-thesis, or #needs-source to track argumentative weight.

Dimension 3: Specific concept (object tags)

This is where your subject-matter tags live, and they should be the most precise tags in your system. Two per note is usually enough. One slightly broader concept, one narrow. For example: #cognitive-bias + #anchoring-effect. The broader tag gives you the neighborhood; the narrow tag gives you the address.

Tiago Forte’s recommendation in Building a Second Brain adds another useful angle here: tag for the project, not the topic. #article-sleep-decisions is more actionable than #neuroscience. Projects have deadlines. Topics don’t.

How many tags per note?

Research and practitioner experience converge on 3-5. Fewer than three and your notes become isolated; retrieval depends entirely on full-text search. More than five and you’re generating noise. The 5-tag rule of thumb, widely cited in PKM communities, recommends a deliberate blend: one or two structural tags (type, status) and two or three subject tags (specific concepts or project affiliations).

Tips that keep the system working

  • Tag for your future search, not your current category. Before tagging, pause and ask: “What would I type into the search bar when I need this note?” Tag for that query. If you’d search “anchoring bias negotiation,” then #anchoring-bias and #negotiation-tactics are your tags. Not #psychology.
  • Keep a master tag list. A single note listing every active tag with a one-line definition. Review it monthly. This is your controlled vocabulary, the term used in library science for a curated list ensuring one concept always maps to one term. Without it, you’ll end up with #cognitive-bias, #cognitive-biases, #thinking-errors, and #mental-shortcuts all pointing to the same thing.
  • Tag at two different moments. At capture, apply one or two quick tags (note type and maybe one concept). During your weekly review, add richer tags based on how the note connects to your current work. Forte calls this progressive summarization applied to metadata. It keeps capture fast and tagging thoughtful.
  • Use relational tags, not just noun tags. Most tags name categories. But tags like #contradicts, #supports, #example-of, or #extends encode how a note connects to your thinking. A note tagged #contradicts + #spaced-repetition tells you something different from one tagged #supports + #spaced-repetition. When you’re building an argument, this distinction is enormously useful.
  • Add one sentence about why you tagged it. Craik and Lockhart’s level of processing research (1972) showed that deeper engagement at the moment of encoding produces stronger, more durable memories. Slapping a tag on a note is shallow processing. Writing “Tagged #anchoring-bias because this study provides a counterexample to Kahneman’s original finding” transforms tagging from filing into thinking.
  • Apply the rule of 2 and 50. If a tag applies to only one or two notes in your entire archive, it’s too specific to earn its keep. If it applies to more than about fifty, it’s too broad to be useful. The productive range is roughly 5-30 notes per tag. Check this during your monthly audit.
  • Let full-text search handle the rest. Forte operated without tags for nearly a decade in an 8,000-note archive before hitting real limitations. If a keyword search would find the note, a tag is overhead. Tags add the most value when your collection exceeds several hundred notes in a domain, when you need to find material across multiple topics at once, or when you’re tracking workflow status.

What to avoid

  • Don’t tag with broad domains. Tagging every psychology-related note #psychology recreates the exact problem folders were supposed to solve. You get one giant bucket instead of one giant folder. The result is the same: you can’t find what you need.
  • Don’t build elaborate taxonomies before you have notes. The Zettelkasten community calls this a systemic failure. Spending hours designing a perfect hierarchical tag tree before you’ve captured anything means you’re organizing hypothetical knowledge, not real knowledge. Start with 8-15 tags. Add new ones only when you have a demonstrated a need for them.
  • Don’t nest tags deeper than two levels. #science/physics/quantum/entanglement looks tidy in theory. In practice, it adds retrieval friction and creates confusion about which level to search. One or two levels of hierarchy is the practical maximum. Beyond that, use links between notes instead.
  • Don’t let AI auto-tag without review. The generation effect, a well-established finding in memory research, shows that information you produce yourself is retained better than information you receive passively. Choosing your own tags forces you to think about what the note represents. AI suggestions are fine as starting points, but the final decision should be yours.
  • Don’t use inconsistent formatting. Pick one convention and stick to it. #cognitive-bias and #CognitiveBias and #cognitive_bias are three tags in most systems, not one. Singular or plural: choose one (singular is the standard recommendation). With or without hyphens: choose one. Case convention: choose one. This sounds trivial until you’re searching for a tag you created eighteen months ago and can’t remember which style you used.
  • Don’t skip the monthly audit. Tags drift. Synonyms accumulate. Old project tags linger. Without periodic cleanup (what PKM practitioners call “tag gardening”), your system degrades quietly until the moment you need it most. Set a recurring calendar event. Go through your tag list. Merge duplicates. Retire anything unused for six months. This takes twenty minutes and saves hours of frustrated searching later.
  • Don’t tag what you should link. Tags and links serve different cognitive functions. Tags describe properties of a note: its type, its status, its concept category. Links describe relationships between notes. If two notes are directly related, link them. Don’t try to force that connection through a shared tag. Tags are for filing. Links are for thinking.

The bottom line

The best tagging system isn’t the most elaborate one. It’s the one you maintain. A small, consistent set of tags that you review monthly will outperform an elaborate taxonomy that you abandon by week three. Start with type tags, status tags, and a handful of precise concept tags. Add new tags only when you catch yourself unable to find something. And remember that the real work isn’t in the tagging itself: it’s in the brief moment of thought about why this note matters and when you’ll need it again.

Resources