Notion for Thesis Management: Hype and Reality

Lina Ortega, a second-year doctoral student in public health at a university in Barcelona, set up her thesis workspace in Notion the week she passed her qualifying exam. It was a beautiful workspace. A linked database of papers with custom tags for methodology and field. A kanban board for chapter progress. A reading log. A timeline view of her committee meetings. An embedded Google Calendar. A nested structure of sub-pages for each research question. She showed it to her cohort over coffee, and three of them set up similar workspaces within a week. By her fourth year, when she was trying to compile her literature review, the main database contained 1,400 entries and took eleven seconds to load on her office wifi. She began to resent the software.

Notion, launched by Ivan Zhao and Simon Last in 2016 after a near-death rewrite of an earlier version, has become the default workspace tool for a cohort of knowledge workers who came of age after Google Docs and before Obsidian. It raised at a ten-billion-dollar valuation in 2021. It acquired the calendar company Cron in 2022, integrated AI features aggressively starting in 2023, and by 2025 had become a genuine competitor to Microsoft and Atlassian for parts of the team-productivity market. Among graduate students, it achieved something closer to cultural dominance. For a period of roughly three years, if you saw a grad student’s laptop over their shoulder in a library, there was a good chance the open tab said notion.so.

The appeal for thesis work is real and worth stating clearly before the critique. A doctoral project is, among other things, a data-management problem. You have papers to track, arguments to organize, deadlines to meet, drafts to version, feedback to absorb, and a sprawling bibliography to maintain over several years. Notion’s linked databases are genuinely useful for this. You can filter your reading list by “cited in Chapter 3” or “flagged for the literature review” with a few clicks. You can link a paper to the argument it supports and pull that linkage into a view next to the argument. The structure is more flexible than Word and Excel combined, and for a certain kind of thinker, building the structure is itself a form of clarifying the project.

What Notion does well is therefore not in dispute. Linked databases, flexible page hierarchies, embedded multimedia, collaborative comments, reasonable templating. For a grant-writing team at a lab, or a small research group coordinating a systematic review, it is a sensible default. Literature reviews in particular benefit from the kind of structured tagging Notion enables. The problems begin when it becomes the only tool, and they compound over the years a dissertation takes.

Performance is the first and most concrete. Notion is a web application. Even in its desktop wrapper, it loads pages over the network, renders blocks client-side, and slows down noticeably as individual databases grow past a few thousand rows. By the end of a PhD program, a reading database can easily contain several thousand entries, and the same is true of notes in long-running projects. Ortega’s eleven-second load time is not atypical. The Notion engineering team has written publicly about the challenge; the fundamental architecture was not designed for power users with a decade of accreted content. Performance improves with each release, but it has not caught up to the expectations of the most demanding users.

Reference management is the second issue and the more specific one. Notion does not integrate natively with Zotero or any other reference manager that handles citations the way an academic actually needs them. You can paste in a DOI. You can build a custom database with fields for author, year, and journal. You can even embed a Zotero export. What you cannot do is cite a source in a paragraph and have the citation update automatically when you change the source record, which is table stakes for serious academic writing. The workflow most grad students end up with involves Notion for organization, Zotero for the bibliography, and Word or LaTeX for the actual chapter draft, with all three sources of truth drifting out of sync over months.

The third issue is the lock-in, and it is more political than technical. Notion’s export format is proprietary. You can export to Markdown and CSV, but nested databases, formulas, and linked relations do not survive the export cleanly. A researcher who has spent four years building a carefully structured workspace discovers, when trying to move out, that the structure was an artifact of the software as much as of the research. This is a contrast that comparative pieces on note-taking systems sometimes elide. Plain text has a portability guarantee. Notion does not.

The honest alternatives are more boring than the Notion workspace tours suggest. A reading list in Zotero, with tags and collections, handles the bibliography correctly and will continue to work in 2035. A folder of Markdown files in Obsidian or plain BBEdit handles the chapter drafts and gives you a version-control-friendly format. A physical or digital kanban handles the task tracking. A shared Google Doc or a proper LaTeX project handles collaborative draft review. None of these is as visually satisfying as a unified Notion workspace. All of them are more robust over the timescale of a dissertation.

The version of the grad-student user that gets the most out of Notion treats it as a capture tool and a project-management layer, not as a writing environment or a reference manager. Ortega’s second workspace, built after her near-breakdown in year four, is closer to this. She uses Notion for her meeting notes with advisors, her committee deadlines, her conference submission tracking, and a high-level map of her chapters. Her actual reading lives in Zotero. Her actual drafts live in LaTeX. Her actual notes on papers live in Markdown files linked to Zotero through a plugin. When Notion is slow, she closes it. The thesis continues.

There is a broader pattern here worth naming. Every generation of productivity software produces a tool that feels, for a few years, like it can hold everything a serious thinker needs to do. Evernote felt this way in 2012. Roam felt this way in 2020. Notion felt this way from roughly 2019 through 2023. The attraction is the promise of a single surface for the entire mess of a long project. The failure mode is always the same: the single surface gets big, slow, and brittle, and the most important parts of the work get trapped inside a container that does not serve them well.

The grad students who finish their dissertations on time are, on the whole, less reliant on any particular tool than the ones who spend weekends customizing their workspaces. This is not a new observation; it is roughly the same observation people made about index cards in the 1960s and about Microsoft Word in the 1990s. The tool is in service of the work, not the other way around. Notion is an extraordinary tool. It is also, for thesis-length projects, a container that will probably not outlast the project it was meant to hold.

The question to ask before building an elaborate workspace is whether the structure will still serve you in year five. If the answer is uncertain, keep the structure light, keep the critical artifacts — drafts, bibliography, data — in formats you own, and use Notion for what it’s genuinely good at. The thesis is the output. Everything else is scaffolding.

Photo via Unsplash.