Starting a thesis means turning a topic or an assignment into a project you can finish. In the first weeks you look at finished theses in your field, narrow the topic, read with a purpose, form a working question, agree on a plan with your advisor, and start drafting.
No one sees their own blind spots, with or without AI help. Esmond points them out in your draft from within your field, along with factual errors and inconsistencies.
Programs order these steps differently and give them different names. Some ask for a thesis proposal within weeks; others let you work toward a first chapter. The table shows what each step should leave you with.
| Step | What you have at the end | Who is involved |
|---|---|---|
| Look at finished theses | A sense of the chapters, length, and conventions in your field | You, your advisor, a subject librarian |
| Narrow the topic | A topic you can state in one sentence, with a group, setting, or period | You, then your advisor |
| Read with a purpose | A source list with notes on what each source argues and leaves open | You, a librarian |
| Form a working question | One question you can answer with data or sources you can get | You and your advisor |
| Agree on a plan | Deadlines, a meeting rhythm, and the approvals you need, in a calendar | Your advisor, sometimes a committee |
| Start drafting | A first chapter or section in a shape someone else can read | You, then your advisor |
A hypothetical example of a commentYour introduction says the thesis is about remote work and well-being, but the methods section describes a survey of one company's IT staff only. Does your research question cover remote workers in general, or this group?
A first comment on one part of your own draft, right away and without an account.
Get a free previewBefore choosing chapters or a method, look at what a finished thesis in your program looks like. The University of Kansas writing center puts it first: “Start by asking your adviser if they have any suggestions of dissertations or theses you should look at.” The UCLA Graduate Writing Center, in a workshop on starting a master's thesis, gives similar advice for choosing a structure: “Review examples of prior master’s theses from your lab/advisor/department/field”.
Read two or three of them for form as much as content. Note the chapter list, how long the literature review is compared with the rest, how the method is described, and where the question appears. For the common forms, see the pages on master's thesis structure and dissertation chapters.
A topic that is still broad makes the start take longer. The economics department at Reed College puts it this way in its thesis guide: “Specific, focused topics are almost always better than highly general or vague ones.” Narrowing usually means adding limits: a group of people, a place, a period, one outcome, or one data source. "Social media and politics" becomes "how first-time voters in one state used short videos during one election."
Reading is how you find out whether the narrowed topic still has something open. The Kansas guide suggests keeping track from the start: “In the beginning you may want to compile at least a list of 10 sources and to organize them in a matrix or an annotated bibliography where you note the subject matter of the source, its novel argument, its theoretical relationship to the subject and field, and what may be lacking within the analysis.” The last point, what each source leaves open, is what you will use later to show your research gap. Put the sources in a citation manager from the start, so the reference list is not a separate job at the end.
A working question is the version you can start with, before it is final. At Bates College, the politics department's guidance on the senior thesis says: “The best theses involve consultation with your advisor before the semester starts in order to identify your research question.”
For tests you can run on your question, see the page on the research question. One check belongs at the start: whether you can get the data or sources the question needs. The Reed guide notes: “If you are doing a thesis that requires empirical data, one of your biggest obstacles is likely to be assembling your data base.” If a question depends on a dataset, an archive, an organization, or interviews, find out in the first weeks whether you have access, and how long it takes to get it.
Prepare the first meetings with your advisor. The Kentucky State University writing center lists, among the first steps of a dissertation: “Create a schedule and set deadlines with your advisor.” The author of a guide for data science master's students at the University of Helsinki adds one more thing to settle early: “agree early on with your thesis supervisor how the feedback on the thesis will work”.
Bring something to each meeting: a one-sentence topic, a short source list, a working question, and the questions you want answered. Afterward, send a short summary of what you agreed. Points worth settling in the first or second meeting:
Plan backward from the submission date, and put the steps that depend on other people in first: approvals, data access, and your advisor's reading time. Some programs tie these together. In the Master of Science in Engineering Management at Western Kentucky University, the thesis guide says that students should not start survey data collection or experimentation for the thesis until the committee has approved the proposal. The Kansas guide adds that if you collect your own data, you should complete the required protocols so the project can be finished within the intended timeframe.
Put these in the plan before anything else:
The UCLA workshop's advice for a first draft fits the start of a thesis too: “Start earlier than you think you should”. Drafting early shows where the topic is still vague and gives your advisor something concrete to respond to. The same workshop suggests beginning from an outline: “Create a preliminary outline of the sections (and sub-sections) you might include. Then, review it with your advisor.”
Which part to start with depends on your program. In some, the opening chapters are the first assignment; the Western Kentucky guide says “The proposal is generally considered as the first three chapters of the thesis.” Where there is no such rule, start with the part you have material for. The method can be described as soon as the design is settled. The literature review grows out of your reading notes. The introduction can start as a working version, but it usually changes once the results are in; the page on the thesis introduction explains when to finish it.
The Helsinki guide gives a reason to finish one or two chapters early on: “it is wise to finish at least one or two chapters early on (e.g., Introduction and Background), and share them with your supervisor so that they can point out major issues before they are all over the thesis”. When you send that chapter, say what you want comments on. The Kentucky State handout puts it this way: “Ask readers to give feedback on specific areas or specific elements that you know need growth.” For who can read a draft before your advisor, see the page on feedback before your advisor reads a draft.
This example is made up; the student, program, and data are fictional. A master's student in public health has the assigned topic "sleep in teenagers." Her first weeks:
Esmond evaluates every assignment before you submit. It never writes it.
Esmond is an AI tool that doesn't write for you. If your course doesn't allow any AI tools, don't use Esmond for that work. If you're unsure, ask your instructor or advisor.
The first point where Esmond is useful in a thesis is the first draft chapter, before your advisor reads it. You upload that chapter, or the proposal if your program starts with one, and Esmond reads it against the standards of your field. It flags factual errors, inconsistencies, and blind spots, such as a question the planned method can't answer or a literature section that doesn't lead to your question, and it points out what your advisor or committee may question. It adds no text to your draft. See how Esmond works and a sample report on a fictional thesis.
Esmond doesn't choose your topic, find sources, or plan your weeks, and it doesn't replace the meetings with your advisor. The steps on this page stay yours; Esmond reads what they produce once there is a draft.
No. A thesis statement is the one-sentence claim at the start of an essay. A thesis in the sense of this page is a research project with a question, sources or data, and several chapters, usually at the end of a bachelor's or master's program or for a doctorate.
Check first whether your program asks for the opening chapters as a proposal. If it doesn't, start with the part you have material for, often the method or the literature review. The introduction usually changes at the end, so treat an early version as a working draft.
Usually there is no need to wait. A working question is enough to draft a methods section or a first literature section, and drafting often shows where the question is still vague. Tell your advisor about larger changes to the question.
It depends on your program's timeline. The Bates College politics department's sample timetable for the senior thesis, for example, asks for a draft of the first chapter on Monday of the fourth week. Your program's own schedule and your advisor decide.
Some programs assign one from a topic form or a proposal, others expect you to approach faculty yourself. Find out which applies to you. In the meantime, a subject librarian or your university's writing center can help with sources and planning.
Read on November 7, 2026.
Get your paper evaluated before you submit