Words, Hours or Pages? Which Writing Metric to Track

A good day’s work can end with a bad number. Three weeks of hard revision, and the word counter reads minus 4,000. The work was real. The unit it was measured in was not.

Anyone who tracks their writing has to choose between words, hours and pages. Most people stay with whatever they tried first and only notice on a day like that one that it has stopped fitting.

What it has to fit is not you but the work in front of you. While a draft is being written, words are the honest unit. During revision it is time. Pages earn their place only when a publisher, a competition or a course has made them the unit. A book passes that switch several times.

The man everyone quotes for word counts ran a clock

Anthony Trollope is the patron saint of the daily word target. Forty-seven novels, written around a job at the Post Office, with a rate he set for himself and hit.

Read the actual passage and the rate is the second thing in it. The first is a watch. He wrote for three hours before breakfast, spent the first half hour rereading yesterday, and only then required 250 words of himself per quarter of an hour (An Autobiography, chapter 15).

Time was the obligation. The word rate only audited it, a check on whether the three hours had been worth anything. Almost every modern retelling keeps the 2,500 words and drops the watch. That is one reason the numbers people quote at each other sit so far above what anyone sustains. The range that holds up is in the article on how many words a day.

That split is the useful part. One metric is the target, the other watches it, and which is which depends on what you are doing today.

A page has never been a unit

Start with the one that confuses people most, because two different things are called a page.

The 250 is arithmetic about paper, not about prose. “60 characters divided by 6 equals 10 words per line. Multiplied by 25 lines equals 250 words per page”, as John Gregory Betancourt spelled it out for SFWA in 2005 (The Obligatory Manuscript Format Article). Six characters per word including the space, ten words to the line, twenty-five lines to the sheet.

William Shunn, whose format guide has been the reference for decades, says plainly why that stopped working. It assumed “a typewriter where ten characters took up exactly an inch and about 25 double-spaced lines fit on a page”. Change any of those factors, he adds, for instance by moving to a proportional font, and “the estimate of 250 words per page no longer holds” (Proper Manuscript Format, April 2021).

Measured against real prose the convention is close but not exact. Counting this blog’s English articles into the same 25-line frame gives 234 words to the page, about 6 per cent under the convention. The full comparison sits in the article on the standard manuscript page.

Where pages are the binding unit anyway

Screenwriting is the clearest case, because the page there is defined by decree rather than by habit.

The Nicholl Fellowships rules for 2026 and 2027 require a feature screenplay “in standard industry format (12 point Courier) and no shorter than 80 pages and no longer than 125 pages”. They add that scripts “outside of standard industry format and/or have cheated margins to exceed 125 pages may have this counted against them during judging” (Nicholl Fellowships rules).

Notice what is actually pinned down. Not the page. The typeface, the point size and the margins. That is the only reason the old rule of one page to one minute of screen time survives at all. Take Courier away and it dissolves.

The lesson generalises. Pages work as a metric exactly when a third party has fixed the format, and not one line before that. Handwriting is the honest exception, because there the page is physical and it converts, though only for your own hand and your own notebook (words per handwritten page).

Why word counts are the default for drafting

A deadline is arithmetic over three terms, the length, the days remaining and the daily quota. Length is stated in words. Hours only enter through your own rate, which is one more estimate in the chain.

That a named number beats a good intention is about as well established as anything in motivation research. Edwin Locke and Gary Latham summarised nearly four decades of it. Specific, difficult goals have been tested “on well over 100 different tasks involving more than 40,000 participants in at least eight countries”. They beat urging people to do their best, with meta-analytic effect sizes from 0.42 to 0.80 (Locke & Latham, American Psychologist 2002).

Their explanation matters more than the number. A do-your-best goal, they write, “has no external referent and thus is defined idiosyncratically”. A word count has that referent and stays checkable on the evening you would rather be generous with yourself.

Turning a date into a quota is then division, and the writing goal calculator does it with the buffer built in.

Where word counts measure the wrong thing

The same paper names the limit, and that half is rarely quoted. On complex tasks, a specific difficult performance goal “sometimes leads to” worse strategies than doing your best, because the pressure produces anxiety. A difficult learning goal, aimed at the process rather than the result, does better there.

Revision is that kind of task. No output of it can be expressed in the unit of the activity. A good pass may add 2,000 words or cut 2,000, and either can be the right call.

Here is a case from this blog, because it is checkable. The English 30-day writing habit plan was reworked on the day it went up, for readability and for four factual corrections. The published page went from 2,726 words to 2,693. A word counter would have logged minus 33 for that afternoon’s work.

Research, outlining and structural surgery all behave the same way. They are work on the text that produces no text.

When the number becomes the target

“When a measure becomes a target, it ceases to be a good measure.” Marilyn Strathern wrote that in 1997, and it now circulates as Goodhart’s law (European Review).

At a desk this is not abstract. Set 1,000 words a day and on a bad day you will write the kind of paragraph that comes easily. A flashback nobody asked for. Dialogue that carries no information. Three adjectives where one would do. The counter climbs. The manuscript does not.

Lisa Ordóñez and colleagues collected the side effects of goal setting into five. A narrow focus that neglects non-goal areas, a rise in unethical behaviour, distorted risk preferences, corrosion of organisational culture and reduced intrinsic motivation. Their proposal is to treat goal setting like a potent medication that needs supervision rather than a harmless tonic (Goals Gone Wild, Academy of Management Perspectives 2009).

The first item describes the novelist who hits the daily target and loses the chapter. The work is organisational research rather than writing research, so what transfers is the mechanism and not a number.

None of this argues against word targets. It argues for tying them to the phase in which the number and the goal still mean the same thing. While you are drafting, they do.

Hours are the unit you actually decide on

You do not decide a word count. You decide whether you sit down at a quarter past nine in the evening. What comes out depends on the scene, the day and things outside your control.

Hours also compare across activities in a way words cannot. An hour is an hour whether you spent it researching, cutting or drafting.

Short sessions are not the fallback option here. Robert Boice taught new faculty to write in brief daily sessions instead of binges. He described the result as effective and comparatively painless (Procrastination, busyness and bingeing, Behaviour Research and Therapy 1989). The study is old, small and has been carrying more weight than it can bear for thirty-five years. The core of it still stands.

The weakness of hours is obvious once stated. They measure attendance. Ninety minutes at the desk with a browser open are ninety minutes, and if you track time alone you pay yourself for sitting.

Which metric belongs to which phase

The third row holds an option people overlook. During revision the natural unit is not the hour, it is the chapter. “14 of 31 done” is honest, it only moves upward, and it does not care whether today’s pass added or removed text.

Two metrics at once

The obvious fix is to count both. Usually that is a mistake, because two targets are two ways to fail on the same evening.

One combination does work, and Trollope was running it. The clock was the commitment. The 250 words per quarter hour were the check on whether the commitment had produced anything.

So one metric is the target and the second is an observation with no threshold attached. Record it, look at it weekly, and do not let it pass or fail your day.

When this does not apply

You are writing with no product. Morning pages, a journal, freewriting. There is no artefact whose length means anything. Hours or a simple tick are enough, and any finer number invents a significance it does not have.

You are paid by length. Then your client’s billing unit is your metric, whether it suits you or not.

You are writing academically. A page of discussion with five checked citations and a page of introduction are not the same work, and no word count represents that. Hours carry better through those phases.

Whatever you end up counting, the tool matters far less than the choice. Paper, a spreadsheet and an app all do the job and differ mainly in how much attention they cost. The five methods compared sets them side by side.

Name the phase your project is in right now. If your current metric is not the one in that row, change it today rather than at the start of the month.

Share