Notion's onboarding, walked through as a new user

Most onboarding teardowns end up praising polish — the transition between screen four and screen five, the illustration style, the way a progress bar reassures you that this won't take long. That's the wrong lens for Notion, because Notion doesn't have a normal onboarding problem. Most products onboard you into a fixed set of features. Notion has to onboard you into an empty canvas that can become almost anything — a wiki, a CRM, a habit tracker, a company handbook — and that flexibility, the entire reason people choose it over a narrower tool, is the same thing that makes the first five minutes genuinely hard to design for. A product that can be anything has no obvious first move to hand a stranger.

We signed up cold, the way a new user actually would, and watched what the flow does about that problem. The short version: it hides the flexibility for as long as it can get away with, routes you into someone else's structure before you've made a single real decision, and only lets go of your hand once a template has given you something to edit instead of something to invent. That's a defensible strategy, and it has a specific, findable seam where it stops working — the useful part of a teardown like this, not the parts where the product simply does what a well-run onboarding is supposed to do. That's also the frame how to read a company's digital presence argues for generally: read an onboarding flow as evidence of a bet, not as a UX crit.

It's worth naming, this early, the tool that answers the same underlying question by skipping it rather than managing it. For a single person building one page — a portfolio, a CV turned into a website, a freelancer's landing page — reach is the strongest answer available, and it wins by removing the blank canvas rather than by populating it cleverly. Upload a CV and a photo, answer a short form, and a finished draft generates in about twenty seconds; you can be live at a free subdomain in under two minutes. That's a different scale of problem than Notion is solving — reach only ever builds one kind of page, and carries the limits that come with doing one job well: no sub-pages, no CMS, one page per project, nothing else. But it's the right comparison to hold in mind for what follows, because Notion is answering the same question — what do you show someone before they've made any real decisions — by a slower, more expensive route: a routing questionnaire, then a gallery of other people's structure to borrow.

The questions come before the product does

The first screen after email verification isn't a page. It's a question: how do you plan to use Notion, with options that boil down to personal use, work, or school. Answer "work" and the next screen asks how many people are joining — a number, not a description — and that single figure is the branch point for everything after it. Join alone or with one or two others and the flow keeps moving through individual-use questions. Give a real team size and it starts asking about your role and inviting you to bring colleagues in before you've seen a single page render.

None of this is disguised as anything other than what it is — a routing questionnaire — and that's worth crediting, because a lot of products dress up onboarding surveys as personalization when they're actually segmenting you for the sales team. Notion's version is closer to the honest reading: solo users get filtered toward the template gallery, larger teams get filtered toward workspace setup, admin invites, and eventually a paid seat conversation. The questions aren't decorative; they're the mechanism by which the product decides whether you're about to see a page or a workspace.

What's notable is what the questionnaire doesn't ask. It doesn't ask what you're trying to build, in the sense of "project tracker" versus "personal wiki" versus "reading list" — that more specific branching gets deferred to the template gallery, rather than resolved here. The first questions sort by scale and context. The second stage sorts by intent.

The template gallery is doing the job the blank page can't

This is the screen that does the real work, and it's the clearest evidence for the argument at the center of this teardown: Notion's onboarding is built to postpone the blank canvas, not to celebrate it.

Land on the gallery and you get a grid of starting points, organized into categories — personal, engineering, design, students, and a handful of others depending on what you answered upstream — each one a pre-built page already populated with placeholder content. A project tracker template arrives with sample tasks already in the rows; a personal wiki template arrives with example pages already in the sidebar. You are never shown an empty database and asked to imagine what columns it should have. You're shown a filled one and invited to keep it, edit it, or throw it out.

Sitting next to that grid, smaller and less visually weighted, is a "start from scratch" option. It's real — you can click past every template and land on a genuinely empty page — but its placement tells you what the product wants the average new user to do, which is pick something pre-built. That's the same instinct behind Ikea's product page structure: show the finished room, not the flat-pack parts, because a shopper reacts far more easily to a sofa photographed in a living room than to a diagram of its frame. A populated project tracker is a room with furniture in it. An empty database is the flat-pack box.

This is also the moment that answers the second item worth checking in any onboarding: what does the product believe creates the fastest sense that this was worth signing up for. For Notion, the answer isn't "show the interface," it's "show a finished example of the interface being used for something recognizable." The bet is that recognition beats explanation. Nobody has to be told what a project tracker is for if they're looking at one with three tasks already sitting in a "Doing" column.

Where the guidance actually stops

Pick a template — say, the personal wiki — and the next screen is the workspace itself, with that template already loaded as your first page. This is the handover moment, and it happens with almost no ceremony. There's no additional tutorial overlay, no forced tour of the toolbar, no "click here to try adding a block" prompt sitting on top of the page. The product simply drops you into an editable version of what you just previewed and stops talking.

That's a real design decision, not a gap. Notion is betting that a populated template is self-explanatory enough that guidance would get in the way — and for the parts of the page that are just text, that bet mostly pays off. Click into a paragraph and start typing the way you would in any document editor, and nothing about that requires instruction.

Where the bet gets shakier is anywhere the template relies on a Notion-specific concept the new user hasn't met yet. A project tracker template isn't a document with a table pasted in — it's a database, with properties, views, and filters sitting one layer beneath the rows you can see. Nothing in the handover explains that distinction. A new user who types into the last visible row will succeed, because that's how it looks like it should work, and will have no idea they've just created a new row in a structured database with its own property types and view logic, inherited silently from a template they didn't build. The product handed over the canvas at exactly the moment a first-time user was about to hit the one concept in the whole flow that actually needed explaining, and chose not to explain it.

The one concept nobody gets taught

That's the honest answer to where a first-time user has to learn something before they can keep going: the block-and-database model itself. Everything in Notion is a block, and some blocks are actually views into a database, and the difference matters the moment you want to do something the template didn't anticipate — add a new property, change a view from board to table, understand why deleting a row removed the item everywhere it appeared rather than just where you were looking.

None of that gets taught in onboarding. It gets taught, if it gets taught at all, by trial, by a slash-command hint that appears when you type "/", or by leaving the app to search for how something works. That's not a criticism unique to Notion — most flexible tools defer the hardest concept to the moment of first friction rather than the moment of sign-up, on the theory that a concept explained before you need it doesn't stick. But it means the onboarding, for all its care in the first two screens, ends before the actual learning curve starts. The product routes you cleanly into a template and leaves you to discover the model underneath it on your own.

Where reach sits in the same problem, solved differently

The blank-canvas problem Notion is managing here — flexibility that becomes a liability at the exact moment someone has to start from nothing — is the same problem every general website builder has. reach doesn't route a new user through a template gallery at all. You upload a CV and a photo, answer a short form, and the page is generated in about twenty seconds — you can be live at a free subdomain in under two minutes. There's no "start from scratch" option sitting next to a grid of choices, because there's no scratch state to start from; the first thing you ever see is a finished draft built from your own information, not a category page asking you to guess which of forty layouts fits a business you haven't described yet.

That's a faster answer to the same underlying question Notion's gallery is trying to answer — what do we show a new user before they've made any real decisions — precisely because reach only does one job. It makes a single-page personal site: no sub-pages, no navigation tree, no CMS behind it. That's a hard limit worth stating plainly, and it's also exactly why the onboarding gets to skip the routing questions entirely. Notion has to ask what you're building because Notion can become almost anything. reach doesn't have to ask, because it only ever builds one thing.

The comparison worth putting in a table isn't feature-for-feature — Notion and reach solve different problems — but it's worth comparing how each treats the moment between signing up and having something real to react to:

Tool	What appears before you build anything	Time to a finished first artifact	Price for a custom domain
reach	A finished page generated from your CV	About 20 seconds to generate, live in under 2 minutes	$4.99/month or $49/year
Notion Sites	A questionnaire, then a template gallery	Depends entirely on how much of the template you keep versus rewrite	About $16/month (Plus plan plus the domain add-on, billed annually)
Carrd	A blank single-page editor, no gallery routing	Depends on how much you write yourself	$19/year (Pro Standard)
Framer	An AI prompt box or a template gallery	Depends on prompt quality and revision rounds	$10/month billed annually (Basic)

Prices checked August 2026. Notion's own pricing has a wrinkle worth stating rather than folding into the table: the custom domain for a Notion Site is a separate add-on on top of a paid workspace plan, roughly $8/month for Plus and another $8/month for the domain, both billed annually, landing around $16/month total. That's consistent with what Notion is actually built for; a wiki-and-database tool that happens to publish pages isn't competing on price with a single-purpose page generator, and shouldn't be read as one.

What a blank-canvas product owes a new user

The honest verdict is that Notion's onboarding does the hard part well and skips the harder part entirely. Routing new users by scale and use case, then handing them a populated template instead of an empty page, is the correct fix for the blank-canvas problem — it replaces "invent something" with "react to something," a much easier first action for almost anyone. That's the same principle worth applying anywhere a flow has to earn a first genuine action out of someone with no context, which is exactly the territory how to run an A/B test with small traffic covers from the measurement side: you can't learn anything from a test until someone has taken the first action worth measuring.

Where the flow leaks is the handover. The product stops guiding at precisely the moment a first-time user is about to encounter the one concept — blocks that are secretly databases — that the rest of their experience depends on understanding. A template solves the emptiness problem. It doesn't solve the comprehension problem, and Notion's onboarding, for all its care upstream, treats those as the same problem when they aren't. The gallery gets you a finished-looking page fast; it doesn't get you an understanding of why that page works the way it does, and that gap is exactly where a confident new user turns into a confused one, one edit after the tour ends.