A thesis is one argument
and a very long calendar.

Write the thesis statement as one sentence: One sentence, in the form of a claim that could be disagreed with. If it takes a paragraph, you have a topic rather than an argument, and a topic cannot tell you what belongs in the thesis and what does not. Put that sentence at the centre of the canvas and leave it visible.

Test every chapter against it: For each chapter, say in one line what it contributes to the claim. Draw the connection. A chapter that needs three lines to justify its place, or that connects to nothing, is usually a good paper that belongs somewhere other than this thesis.

Reorder until the through line reads: Most argument problems are order problems. Cards move with a drag, so you can try the methodology chapter before the second analysis chapter and read the resulting spine immediately, rather than restructuring a document to find out it did not help.

Watch for the four loosely related papers: This is the most common structural failure in a thesis, and it is almost undetectable from inside chapter three. Read the chapter cards left to right with nothing else on screen. If they do not build, you have found it while there is still time to fix it.

What each chapter owes the argument: The introduction states the claim and the stakes. The literature review establishes the gap the claim fills. The methodology justifies why your approach can answer the question. The results or analysis chapters produce the evidence. The discussion interprets it against the field. The conclusion states what is now known that was not known before.

The literature review is its own layer: Reading, synthesis, and finding the gap are a large enough job to deserve their own map. Plan the thesis here and run the reading in the literature review organizer, on the same canvas, so the sources feed the chapters that depend on them.

Methodology and data as a visible chain: Connect the research question to the method, the method to the data it produces, and the data to the analysis chapter it feeds. A break in that chain is the question a viva examiner will find, and it is much cheaper to find it yourself in year one.

Write the introduction last: The introduction has to promise exactly what the thesis delivers, and you do not reliably know what it delivers until it is written. Sketch a working introduction early so you have a direction, then write the real one at the end, when the promise can be accurate.

Work backwards from submission: Put the real deadline on the timeline first, then the revision round before it, the full draft before that, and each chapter draft before that. Planning forwards from today produces a schedule that ends whenever it ends. Planning backwards produces one that has to fit.

Supervisor cycles always take longer than planned: A chapter goes out, and it comes back in three to six weeks, not the ten days the plan assumed. Put every review cycle on the timeline as a real block of elapsed time, and write another chapter while one is out. Two supervisors means two calendars to reconcile.

Build slack for ethics, recruitment, and data: Ethics approval, participant recruitment, archive access, lab time, and transcription are the classic overruns, and they are the parts you do not control. Mark them on the timeline with explicit slack rather than assuming best case, because the best case is not what happens.

Schedule the revision round as real work: The gap between a full draft and a submittable thesis is months, not a tidy-up. Formatting, restructuring after supervisor comments, the abstract, and the final read all belong on the calendar as their own phase, and leaving them off is how a plan that looked fine runs three months late.

Set a weekly target you will actually hit: Most people plan at a pace they have never written at. Track a few real weeks first, then set the target from the measured pace rather than the hoped one. A schedule built on five hundred words a day that you have never once produced is a schedule that fails in month two.

Keep the planned and actual lines both visible: Mark what you planned to have written by each point and what you have actually written. Seeing the gap early makes it a small correction. Seeing it late makes it a crisis, and the whole value of a long timeline is that it is early for a very long time.

Share a view-only link with your supervisor: A supervisor who can see the argument map and the timeline at once gives more useful structural feedback than one reading a chapter in isolation. They open it in a browser with no account. If they are on a paid plan themselves, anyone they invite to a board joins free.

AI that reads the board, not the thesis: Because the AI has the whole canvas as context, you can ask it to draft a chapter structure, propose a timeline, or point out which chapter has no visible link to the claim. It is there to help you structure the project. The argument, the analysis, and the writing stay yours.