How do I write consistently during a PhD?

Most doctoral writing advice assumes the problem is discipline. Usually it is scheduling and scale. Waiting for a clear day means writing on the few days that are clear, which in a research degree is perhaps one week a month, and a thesis does not get written in twelve weeks spread over three years.

The alternative is a small daily commitment that fits the days you actually have. Ninety minutes at the same time each morning, before teaching, before email, before the day gets claimed. The size matters because it has to survive a bad week, and a two-hour target is abandoned in a bad week while a forty-minute one is not.

The published advice says the same thing and gives a reason beyond output. Peterson and colleagues argue that writing every day, even briefly, creates time for reflection, lets ideas mature, and improves scientific thinking rather than only recording it (Peterson et al., 2018, pp. 1-2). They pair that with a concrete setup: choose a location that limits distraction, put unrelated tasks on a list to handle later, and turn off email and internet access for the session, because interruptions both break the thinking and eat the time (Peterson et al., 2018, p. 2).

Protect the start rather than the whole session. Most of the resistance is in the first five minutes, so remove decisions from that window: know before you sit down which section you are working on and what the first sentence is doing. Ending a session mid-paragraph, with a note about what comes next, makes the next start almost free.

Track sessions rather than words. Word counts fluctuate for reasons that have nothing to do with effort, and a day spent cutting two thousand words can be the most productive of the month. A simple record of whether you sat down is a better signal and a kinder one.

Not everyone agrees that counting should be dropped. Daudt and colleagues, reporting on student-led writing retreats, ask participants to arrive with a stated goal, either quantitative such as a word or page count, or qualitative such as revising a named section, and they found that sharing the goal in advance removed decisions during the session and showed people whether their target was realistic (Daudt et al., 2026, p. 4). Participants in those five-day retreats averaged 1.2 pages, about 550 words, per day. If you do set a number, that is the scale of a working figure, not two thousand words before lunch.

How long should a writing session be?

Long enough to get past the reorientation, short enough that you will start. For most people that is 45 to 90 minutes. Sessions past about two hours produce text you rewrite the next day, which is why total hours is a poor measure of writing progress.

If you have a whole day, use it as two or three sessions with real breaks rather than one long block. The output is higher and the recovery cost is much lower.

The retreat timetable is a usable template for a full day. Daudt and colleagues ran roughly 9:00 a.m. to 6:00 p.m., built from writing blocks of 60 to 120 minutes with breaks of 30 to 60 minutes between them, a 15-minute low-pressure warm-up first thing in the morning, and a review with an accountability partner in the evening (Daudt et al., 2026, p. 4). Treat that as a reported account of what those students did rather than proof that the same schedule produces the same output for you (Daudt et al., 2026, p. 7).

Separate drafting from editing into different sessions. They use different attention, and switching between them within a session is the most common reason a writing hour produces one heavily polished paragraph.

What do I do on days when nothing comes?

Lower the target rather than skipping. Write badly, write the outline, write a paragraph about why the section is stuck, or revise something already drafted. The habit is the asset. A day producing 80 useless words keeps the sequence intact and tomorrow starts easier.

Writing about the block is genuinely productive more often than it sounds. "This section will not move because I do not know whether the second study belongs here" is a diagnosis, and it is usually the actual obstacle rather than a lack of motivation.

If nothing comes for a fortnight rather than a day, that is a different problem. It usually means a structural decision has not been made, or something outside the work is taking the attention, and neither responds to trying harder at the same hour.

Missed days are expected, and the useful response is to read them rather than punish yourself for them. Peterson and colleagues recommend an external commitment or incentive that makes designated writing time hard to cancel (Peterson et al., 2018, p. 3), and marking writing on a calendar so that gaps become visible, while stating plainly that writing practices have cyclical ups and downs (Peterson et al., 2018, pp. 4-5). When resistance shows up, their advice is to name what is causing it and change the strategy instead of dropping the practice. For a stall that lasts weeks in the final year, Marino and colleagues recommend agreeing an action plan with your supervisor about a year before submission, reviewing progress every few weeks, and building in contingencies for technical failures (Marino et al., 2014, p. 1). They deliberately do not prescribe one planning method, since a timeline, a mind map, or a stepwise sequence can each work as long as the plan stays flexible without turning into another form of procrastination.