Writing Over the Long Run: Managing Years of Output

The Compounding Asset

After a year of serious writing, you have a corpus: a body of work. Memos, docs, notes, PR descriptions, journal entries, essays. Hundreds or thousands of files.

After five years, you have a library. Not in the sense of books; in the sense of a reference set. Your past thinking, documented.

This corpus has real value:

Pattern recognition: your past thinking on a topic informs your current thinking
Reuse: templates, framings, and arguments get lifted from old work
Teaching: you have real examples to hand to new hires or collaborators
Self-knowledge: reading old notes reveals your own patterns, for better and worse

The trick is making the corpus usable. A disorganised 5000-file mess is theoretically valuable and practically inaccessible.

Search Over Your Own Corpus

The single most important infrastructure: you can search your own writing.

Plain text in a folder: grep, ripgrep, or your editor's search
Notion, Obsidian, and similar: built-in search
Google Docs: Google's search
A mix: whatever catches all your writing

Why search matters

Most of your writing's value is retrospective. You won't know what piece of writing matters until six months later when you remember writing "something about onboarding" and want to find it.

Search compensates for the impossibility of pre-organising the corpus. You don't need to know where you put the file; you need to be able to find "that thing about X" when it matters.

Making writing searchable

Specific titles: not "Meeting notes" but "2026-03-15 product sync"
Keywords in the body: if you're talking about authentication, use the word "authentication" (not just "auth")
Tags if your system supports them: don't over-invest, but basic categorisation helps
Consistent file naming: dates-first, topic in the name

A lot of writing is un-findable because nobody thought about findability at write time. Fix that habit once; pay for a decade.

The Tagging Question

Tags and metadata are useful but easy to over-invest in.

When tags help

You have a clear small set (5-20 tags)
The tags correspond to your actual information needs
Each piece gets 1-3 tags, not 10

When tags don't help

The tag system is more elaborate than the content
Adding tags is the main activity
You can't remember the tag scheme

A rough rule: if tagging is slowing down capture, simplify. If you can't find things without tags, invest in tags. Match the overhead to the actual payoff.

Folksonomy vs taxonomy

Folksonomy: tag as you go, emergent structure. More natural; messier over time
Taxonomy: pre-defined categories. Cleaner; often doesn't match what actually comes up

Most personal systems work better as folksonomy. You learn what categories matter by the tags you keep reaching for. Pre-defining produces structures that don't serve the content.

Templates Over Time

Over time, you accumulate working patterns: the memo shape that produces good decisions, the PR template that reviewers like, the ticket format that gets acted on.

Codifying patterns

Turn working patterns into explicit templates. Not just in your head; written down:

A template file you copy from
A template system built into your tool
A cheat sheet of reusable openings, transitions, closings

When you find yourself writing "the same shape" a third time, make a template. By the fifth time it would have repaid the initial setup.

Templates at the team level

If you're writing on a team, shared templates pay back even more. The team's memos, docs, and tickets get more consistent; new joiners have an easier time.

Shared templates are political (people have opinions). Start with a draft, circulate, iterate, ship. Don't try to design them in committee; design alone, get feedback, revise.

Writing Archives as Research Material

One of the best uses of your old writing: as input to new writing.

Writing a post about X? Search your notes for X. Reuse good paragraphs
Making a new pitch? Lift frames from pitches that worked before
Building a doc for a new hire? You've written versions of this before; start from them

Most writers re-invent work they already did because they forgot they did it. A searchable corpus turns this into: "find the previous version, extract what's reusable, write the new delta".

The Second Brain Movement

A recent meme: Personal Knowledge Management (PKM), "second brain", "digital garden". Tiago Forte's Building a Second Brain is the canonical book.

The core idea: treat your knowledge-capture infrastructure as a serious practice. Capture, organise, distill, express.

What's useful

The emphasis on active engagement (distilling, not just collecting)
The bias toward writing in your own words (processing, not copying)
The habit of periodic review

What's less useful

The tendency to fetishise the system
Elaborate setups that become ends in themselves
The implicit claim that more captured information equals more wisdom

If PKM methodology helps, use it. If the system becomes the work, simplify.

Publishing: The Archive as Output

Some writers turn their corpus into public output over time:

Newsletters built from notes
Blog posts drafted from private journal entries
Books compiled from years of essays
Talks assembled from accumulated arguments

Publishing is one way to put your corpus to work. Not the only way (much of the value is private), but a common one.

Why publish

It forces better writing (a reader cares more than you do)
It creates an accidental community of like-minded readers
It produces feedback you wouldn't get privately
Some of the writing becomes professionally valuable

Why not

Takes energy
Attracts attention you might not want
Introduces performance pressure that can kill private exploration
Most of what you write isn't worth public attention

Publish if it fits. Don't feel obligated. Private writing is complete in itself.

The Long Arc

At year 5 of serious writing, patterns emerge:

You have recurring themes, whether you meant to or not
Your style has settled into something distinctive
Your earliest work looks thin to you (good sign; means you grew)
Your productive zones are clear
Your weak zones are clear too

These are all available through the corpus. They're much harder to access without writing things down and keeping them around.

Most of this only happens if you stayed with a practice. A year of writing, then abandoning it, then another year elsewhere produces a fragmented corpus. A sustained practice in one place produces material you can stand on.

When the Corpus Becomes a Burden

A warning: at some point, the corpus can become heavy.

Signs:

Search returns too many hits to triage
Old documents embarrass you (their wrongness is obvious now)
You feel obligated to maintain things you wrote years ago
New writing feels like it has to justify itself against the archive

Responses:

Aggressive pruning: delete old work you no longer stand behind
Archive partitioning: move older work out of the active workspace
Fresh starts: begin a new notebook, a new folder, a new space. Old corpus still searchable; new work gets breathing room
Permission to change: your current view can differ from your 2019 view. Don't be held hostage by what you wrote then

The corpus is a tool; it should serve you, not the other way around.

Common Pitfalls

"I need a perfect organisation." Perfect organisation is a mirage. Good-enough plus good search beats elaborate structure that no one maintains

"I'll organise the archive when I have time." You won't. Either organise incrementally as you go, or accept that the archive is a search-based index

"Old writing is valuable by default." Some of it is. Much isn't. Pruning is a feature, not a failure

"A bigger corpus is better." A focused, meaningful corpus is better than a huge low-signal one. Size isn't the metric

"If I had the right tool..." You don't. Most corpus problems are discipline problems, not tool problems. Tools help the discipline; they don't substitute

Next Steps

Continue to 12-best-practices.md for the habits that hold the practice together.