Both systems work, and they fix different failures. Stephen King's 2,000 words a day, set out in On Writing (2000), is a word target: it forces output past resistance. Anthony Trollope's 250 words every fifteen minutes, three hours before his postal job — described in his Autobiography (1883) — is a time target with a rate attached: it protects the session itself. A word target punishes slow days; a time target punishes skipped days. The useful question is which failure you actually have.
What does a word target actually train?
Throughput. A daily word count converts the vague intention to write into a number that cannot be argued with, and its documented strength is momentum: King describes the first draft arriving in a season, and Trollope — the proto-case — produced forty-seven novels largely before breakfast. The word target's premise is that quantity is a form of quality control, that a writer producing steadily will fix in revision what speed lets through.
The premise holds only while the draft is moving forward. A word target applied to a structural problem produces pages that will be deleted, written at full daily quota. The number stays honest while the manuscript quietly degrades — which is why experienced writers describe word-target seasons and word-target disasters with equal conviction.
What does a time target protect?
The session. Committing to hours rather than output keeps the practice alive through the stretches where no words are coming: research days, blocked chapters, the week a scene refuses to resolve. Haruki Murakami's documented routine — four to six hours at the desk through a writing season, paired with running, as he has described in interviews including his 2004 Paris Review conversation — treats the hours as fixed and lets the output vary inside them. The time target says the writer shows up regardless; it refuses to reward the day that produced 1,200 disposable words more than the day that produced one true paragraph.
The documented risk is drift. Time without any measure invites the slow elaboration of rituals — tea, re-reading, note-taking — that fill the hours while the draft stops advancing. Trollope's answer was hybrid: the clock ran, but the words were counted within it, at a pace he reported as roughly 250 per quarter hour, and he famously moved on mid-sentence when the session ended so the next morning began moving immediately.
How do you tell which failure is yours?
Audit one honest week. If you write often but stall in revision and generate pages you later cut, the problem is unmanaged speed, and a time container with a modest floor works better. If you protect your writing time perfectly and produce almost nothing in it, the problem is resistance, and a low, non-negotiable word target — small enough to be embarrassing to miss — is the documented cure. The two failures look identical from the outside and require opposite instruments.
| Your pattern | Better fit | Why |
|---|---|---|
| Misses sessions entirely | Time target first | Rebuilds the habit before demanding output |
| Shows up, produces little | Low word target | Forces contact with the sentence-level work |
| Produces heavily, cuts heavily | Time target plus phase rules | Separates drafting speed from structural decisions |
| Fast drafter, avoids revision | Word target on new pages, time target on revision | Revision resists counting; it responds to scheduled hours |
Two cautions from the documented cases. None of the routines ran seven days a week indefinitely — Trollope's hunting seasons and King's stated afternoons off appear in the same accounts as the quotas — and all of them were built around fixed life constraints rather than against them: Trollope's postal job, Murakami's running, King's afternoons. A target that requires an empty calendar is a vacation plan, not a practice.
Related stories: Drafting Software Compared · Writing Retreats: What They Cost.
What about tracking the number itself?
Trollope kept a daily log of words written against time — the Autobiography records the habit — and modern writers replicate it with spreadsheets and apps. Tracking works for the same reason the target works: it makes drift visible within days instead of chapters. The documented failure mode is measurement capture: when the log becomes the product, writers begin writing to the log — padding sentences, re-drafting already counted pages, producing words that exist only in the count. Guard rails keep it honest: track only new prose, log revision hours separately, and audit the log monthly against actual manuscript growth — the word count of the file, not the sum of the daily entries.
A quieter use of the log is diagnostic. Rates fall before stalls happen: a writer whose steady 800 becomes 600 for two weeks is usually avoiding a structural problem three chapters ahead. The writers who get the most from documented routines read the numbers that way — as instrumentation, not judgment.
What did the documented routines share?
Three properties, across otherwise different systems. They were anchored to a time of day — Trollope's 5:30 start, King's mornings, Murakami's 4 a.m. wake-ups during drafting seasons — so the decision to begin was never renegotiated daily. They were modest enough to survive bad days; none of the documented systems depended on inspiration. And they separated drafting from revising absolutely: the target governed new pages only, with revision handled in its own scheduled hours. Writers who merge the two under one target end up counting polished paragraphs against raw ones and burning out on the arithmetic.
How do you switch systems without losing the streak?
Change one variable at a time and pre-commit the rule in writing. Moving from words to hours: set the hours where the words used to be, and keep a private floor — say, one true sentence — so the session cannot end at zero. Moving from hours to words: start the count absurdly low, at a number you hit even while exhausted, and raise it only after two weeks of clean hits. The documented streak-keepers, Trollope most famously, treated the rule as external to mood; the rule, not the day's feeling, decided what happened at the desk.
Whichever instrument you choose, attach it to a phase. Drafting rewards word targets; revision and research reward time targets; stalled projects reward the smallest target of either kind that still forces daily contact. The writers whose routines get published are the ones still writing — which usually means they treated the system as a tool with a job, not as an identity.
Frequently asked questions
Is 2,000 words a day realistic for a working writer?
Not usually alongside another job, and King's number comes with his full-time framing. Trollope hit his totals in three pre-work hours; Murakami treats his seasons as full days. Part-time writers commonly document workable floors between 300 and 800 words, with consistency doing the compounding.
Do streaks matter or are they motivational theater?
The documented streak-keepers used them to remove the daily renegotiation, not to gamify output. A streak that produces defensive minimum pages can protect bad material; a streak that protects the session itself — you sat down, you worked — matches every documented version of the practice.
Should I count hours spent outlining or researching?
Count them inside a time target, never inside a word target. Research and outlining are drafting's front porch, but counting them as pages is how a word target decays into decoration. If a week produces notes only, that is a time-target week, and the word target resumes with the draft.