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.
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:
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:
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
What to avoid
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.
