Obsidian vs Notion vs Roam: Which Personal Knowledge Tool Is Worth the Switch

I have used Notion since 2019, Roam Research from 2020 to 2023, and Obsidian from 2022 through today. At one point, for about eight months, I was paying for all three and actively using all three. This was, with hindsight, a mistake. Each of these tools is genuinely good. None of them are good at the same thing. Running all three in parallel meant every note-taking decision triggered a question about which tool to use, which is exactly the failure mode these tools are supposed to prevent.

Eighteen months ago I consolidated to Obsidian as my primary system, kept Notion for a narrow set of use cases, and stopped paying for Roam. The thinking that led me there is maybe more useful than the recommendation itself, because different thinkers will land in different places, and the right tool depends on how you actually work — not on which one is trending on Twitter.

The Three Tools, Honestly Described

Notion — the database in hipster clothing

Notion's fundamental architecture is a block-based relational database. Everything is a block; blocks can be structured into pages, databases, views. The killer feature is the database: you can build a reading list, a contact list, a project tracker, all with rich fields and multiple views.

Notion's weakness, for deep thinking, is that it encourages structure-first thinking. You spend time setting up the perfect template, the right database schema, the ideal view, and the actual notes are a secondary consideration. It's a tool that rewards people who like organising things.

When Notion is the right answer: managing projects with structured metadata, shared team workspaces, running processes where the same template is used repeatedly, maintaining living reference documents (company wiki, team handbook, client onboarding).

When Notion is the wrong answer: freeform thinking, messy idea capture, anything that needs to happen fast, anything you want to search quickly across your full history.

Roam Research — the original, now aging

Roam pioneered the bidirectional linking paradigm in 2019. The daily notes page plus [[wiki-links]] became the standard pattern for tools-for-thought. Conor White-Sullivan, Roam's founder, genuinely changed how a generation of knowledge workers thought about note-taking.

Roam's problems are structural. It's expensive (around $165/year at the personal tier). It's slow — the app has noticeable lag at even modest note counts. Development has stalled while competitors have accelerated. Your data is locked inside Roam's database; exporting is possible but the resulting markdown loses much of the connection structure.

When Roam is the right answer: if you were an early adopter, got value from it, and moving everything out is more expensive than continuing to pay. Otherwise, rarely.

When Roam is the wrong answer: almost any new user starting today. Obsidian does the same job, faster, cheaper, and with portable data.

Obsidian — local files, bidirectional links

Obsidian stores your notes as plain markdown files in a folder on your computer. It then wraps that folder in an interface with bidirectional linking, graph visualisation, a plugin ecosystem, and an extensive community. The key architectural decision: your data is just files on your disk, editable in any text editor, searchable with any tool, syncable to anywhere.

Obsidian's weaknesses are specific. It's not good for team collaboration — the file-based model makes shared editing awkward. The plugin ecosystem is both a strength and a liability; it's easy to over-configure. The mobile app has historically lagged behind the desktop.

When Obsidian is the right answer: personal knowledge work, Zettelkasten-style note-taking, anyone who wants to own their data, anyone who intends to use the same notes system for a decade. The local-file architecture is the single most important feature for long-term users — it means your notes will still be readable in twenty years, when whatever app you're using today has been acquired, pivoted or shuttered.

When Obsidian is the wrong answer: team knowledge bases, process-heavy project management, situations where the notes are collaborative artefacts.

The Question That Should Drive Your Choice

The most useful diagnostic is not "which features matter to me?" — every tool's marketing makes that confusing. It's "what is my primary failure mode as a note-taker?"

• If your failure mode is accumulating disorganised notes you never revisit: Notion. The structural scaffolding will force some discipline.
• If your failure mode is over-structuring and never actually writing: Obsidian. The loose file structure gets out of your way.
• If your failure mode is losing work when tools change: Obsidian. Markdown files survive any migration.
• If your failure mode is isolation — you want a tool that integrates with how your team already works: Notion. It plays better with others.
• If you're trying to replicate an academic Zettelkasten practice: Obsidian. It's the closest fit to Luhmann's method in software form.

The honest answer for most people: the primary failure mode of note-taking systems is abandonment. You'll use whichever tool you set up for about three months and then stop. The tool that survives is the one with the lowest friction for your actual habits, not the one with the best features.

The Migration Cost Nobody Talks About

Switching between these tools is not cheap. Any note-taking system becomes valuable over months and years; a new tool means starting from zero. Before switching, ask honestly whether the tool is the problem or whether your habit is the problem. Most of the time, the habit is the problem. A new tool gives you a honeymoon of enthusiasm that masks the underlying issue for a few months, and then you're back to the same pattern in a shinier container.

I've watched multiple colleagues cycle through Notion → Roam → Obsidian → Logseq → Obsidian again over five years, never accumulating meaningful value in any of them. The moves were each rational in isolation. The pattern was irrational in aggregate. A mediocre tool used for a decade produces more value than four excellent tools cycled through over the same decade.

What I Actually Use

For full transparency, my current stack, 18 months into stability:

• Obsidian for everything personal. About 3,400 notes at this point, linked, searchable, synced across devices using iCloud. This is my primary knowledge system.
• Notion for three specific things: my consulting client tracker (relational database fits), my reading list with structured metadata, and shared team docs when I collaborate with clients who use Notion.
• A plain text file called scratch.md for anything ephemeral — meeting notes that will be synthesised later, temporary ideas, phone-call notes. Most of these get deleted within a week.

The split maps to the tools' strengths. Obsidian is for thinking. Notion is for structured reference. Plain text is for fleeting capture. Each tool does what it's actually good at.

The Uncomfortable Truth About the Debate

The online discussion of these tools has a quality that would embarrass people if they stepped back. Grown adults arguing about whether Obsidian's block-linking implementation is superior to Roam's, as if this were a meaningful dimension of human life. The tool is not the point. The thinking you do with the tool is the point.

The single most important variable in determining the value of any of these tools over a ten-year horizon is your habit of writing. The difference between someone who writes 500 thoughtful notes a year in Obsidian and someone who writes 500 thoughtful notes a year in Notion is small. The difference between either of them and someone who writes 50 notes a year in the perfect tool is enormous.

Pick one. Commit for at least two years. Do the work. The tool that survives that two-year test is the right tool for you, almost regardless of its comparative strengths and weaknesses at feature level.

The One Recommendation If You're Starting Today

If you're genuinely starting from scratch and don't have an existing commitment to any of these tools, start with Obsidian. It's free for personal use. Your data is yours in a format that will outlast the app. It's the closest software fit to the Zettelkasten method, which is the practice most worth replicating from the academic literature on thinking-through-writing.

You can always migrate out. Markdown files are readable everywhere. The reverse is not true — starting in Notion and trying to migrate five years of work to Obsidian is painful and lossy. Start with the portable option, even if you end up preferring something else. The portability is the insurance policy.

Whatever you choose: start today. Write your first real note in the next 30 minutes. Do not read another comparison article. They'll still be there in three months if you decide to switch. By then, if you've actually been writing, you'll have a real basis for the comparison — which is the only basis that matters.