🧙 UX Pattern Guide

Onboarding Wizard: When to Use One, How to Design It & 6 Patterns That Work

What belongs in a setup wizard, what should be a checklist instead, the seven rules that decide whether people finish it — and six patterns you can copy this week.

📅 Updated July 2026 ⏱ 14 min read ✍️ By Kompassify
An onboarding wizard screen showing a three-step progress indicator and a skip option

Almost every SaaS product has one: the short sequence of screens that appears immediately after signup, asks a handful of questions, and hands the user a configured workspace. Some of them are the reason users reach value on day one. Most are the reason they never get past day one.

The difference is rarely visual design. It is whether the wizard is collecting things the product genuinely cannot work without — or whether it has quietly become the place every team appended their favourite question.

This guide covers what an onboarding wizard is, when it is the right pattern (and when a checklist or product tour would serve better), how to design one that people finish, and six wizard patterns worth copying.

An onboarding wizard compared with an onboarding checklist and a product tour

Three different jobs. Most weak onboarding is a wizard doing a checklist's work.

Key Takeaways

  • A wizard is for setup, not for teaching. It should collect only what the product cannot function without.
  • Every step must earn its place. The test: if this answer were missing, would the product be broken or merely less personalised?
  • Show progress honestly. A visible, accurate step indicator is one of the largest single levers on completion.
  • Let people skip. A skippable step gets abandoned; an unskippable one gets the whole signup abandoned.
  • End inside the product, never on a summary screen. The last click should place the user in a workspace that already has something in it.
  • Instrument every step. Wizard drop-off is the most precisely diagnosable funnel you will ever own.

What is an onboarding wizard?

Onboarding wizard definition: a short, linear sequence of screens shown immediately after signup that collects the information and choices a product needs before it can be useful — account details, workspace configuration, integrations, and team invitations. Unlike a checklist or a tour, a wizard is blocking: the user completes it, skips it, or abandons signup.

The blocking part is what makes wizards powerful and dangerous in equal measure. You have the user's full attention exactly once, with no competing UI. Every question you ask is paid for out of that attention, and the budget is smaller than most teams assume.

Wizard vs checklist vs tour

Onboarding wizardOnboarding checklistProduct tour
ShapeLinear, blockingNon-linear, persistentGuided, contextual
JobConfigure the productDrive a sequence of first winsExplain the interface in place
TimingOnce, at signupAcross the first days or weeksWhen the user reaches a feature
Failure modeToo long, users abandon signupIgnored if items feel like choresClicked through without reading
Right length2–4 steps3–6 items3–5 steps per tour

The most common structural mistake in SaaS onboarding: putting checklist work inside the wizard. "Create your first project", "Connect your calendar", "Explore the dashboard" are all valuable — and none of them belong in a blocking sequence between the user and the product. Move them to a checklist the user can work through at their own pace.


When you should (and should not) use a wizard

✅ Use a wizard when

  • The product genuinely cannot render a useful screen without input (a workspace name, a data source, a currency)
  • Early choices meaningfully branch the experience
  • Setup requires an irreversible or hard-to-change decision
  • An integration must be connected before anything works
  • Your product is multi-tenant and the user must create or join an organisation

❌ Use something else when

  • You are collecting marketing or enrichment data
  • The information could be inferred or defaulted
  • You want to teach the interface — that is a tour's job
  • The task is a first win rather than a prerequisite
  • The answer only changes cosmetics

The question that removes half of most wizards: "If this field were blank, would the product be broken, or just less personalised?" Anything in the second category belongs in settings, in a checklist, or nowhere. Company size and job title are the classic offenders — genuinely useful to your go-to-market team, and pure friction for the user trying to reach value.


How to design an onboarding wizard people finish

  1. Cut the steps to what the product cannot work without

  2. Show accurate progress at every step

  3. Ask one thing per screen

  4. Default everything you can

  5. Make every step skippable

  6. Use early answers to branch what happens next

  7. End inside a workspace that already has something in it

1. Cut the steps to what the product cannot work without

List every field, then delete anything that fails the broken-or-less-personalised test. Two to four steps is the range where completion stays high. Beyond five, drop-off climbs sharply and the steps at the end perform worst — which is unfortunate, because teams tend to put the hardest and most valuable step (connecting an integration, inviting the team) last.

2. Show accurate progress at every step

A step indicator does two jobs: it tells users how much is left, and it commits you publicly to not adding more. Users who can see "step 2 of 3" complete at meaningfully higher rates than users facing an unbounded sequence.

The rule that matters: the indicator must be honest. A wizard that says "step 2 of 3" and then reveals a fourth step destroys trust at exactly the wrong moment. If steps are conditional, count the maximum, not the minimum. The same principles that govern a good onboarding progress bar apply here.

3. Ask one thing per screen

A single question with plenty of space consistently beats a dense form, even though it means more screens. Each screen becomes a trivial decision instead of a small project, and — critically — you learn exactly which question people refuse to answer. A ten-field form that gets abandoned tells you nothing about which field caused it.

4. Default everything you can

Pre-fill from the signup email domain, the browser locale, the timezone, the invite that brought them in. Every field a user does not have to think about is a field that cannot cause abandonment. Sensible defaults with an obvious way to change them beat empty required fields in essentially every case.

5. Make every step skippable

This is the recommendation teams resist hardest, and the data is not ambiguous: a skippable step loses that step, while an unskippable one risks losing the user. Make "Skip for now" visible but secondary, and — the part most teams forget — put the skipped item straight into the onboarding checklist so it is not lost, just deferred.

6. Use early answers to branch what happens next

The strongest reason to have a wizard at all is that it lets you personalise everything downstream. A role or use-case question in step one can select which template is created, which tour runs, and which checklist appears. This is what multi-choice onboarding is for, and it is the difference between a question that costs the user something and a question that pays them back immediately.

If you ask a branching question and then show everyone the same next screen, delete the question.

A branching first step in an onboarding wizard, with the choices on the left and the rendered step on the right

One branching question in step one, and everything downstream can be personalised to the answer.

7. End inside a workspace that already has something in it

The last click of a wizard should not land on "You're all set!" — it should land the user in the product, with their first object created, sample data loaded, or the view they configured already open. An empty state at the end of a wizard wastes the entire investment: the user did the work, and the product responded with nothing.


6 onboarding wizard patterns worth stealing

Pattern 01 · Personalisation The role-branch first step

Step one asks a single question — "What will you use this for?" — with three to five large, visual options. Everything afterwards is shaped by the answer: templates, sample data, the first tour, the checklist.

Why it works: it is the one question users are happy to answer, because the payoff arrives on the very next screen. Keep the options few and concrete; "Other" as a text field kills the branching value.
Pattern 02 · Instant value The pre-populated workspace

Rather than asking users to create their first object, the wizard creates one for them based on their answers — a sample project, a demo dashboard, a template document clearly marked as an example.

Why it works: it removes the blank-canvas problem entirely. Make the sample obviously deletable, or users will hesitate to touch anything in case they break something real.
Pattern 03 · Time to value The single-integration wizard

For products whose value depends on customer data, the entire wizard is one screen: connect a source. No profile questions, no preferences — just the one action that makes everything else possible.

Why it works: it refuses to spend attention on anything but the critical path. Always offer a "explore with sample data" alternative for users who cannot connect production data on a first visit.
Pattern 04 · Progressive disclosure The two-step wizard plus deferred checklist

The wizard collects only the two genuinely blocking items. Everything else that used to be step three through seven moves into a persistent checklist the user works through over the following days.

An onboarding checklist holding the steps deferred out of the onboarding wizard

Everything that is not strictly blocking moves here, where users can work through it at their own pace.

Why it works: it separates prerequisites from progress. This is the single highest-impact restructuring available to most teams with a long wizard.
Pattern 05 · Team products The invite step with a real reason

Instead of a generic "invite your team", the step explains what specifically unlocks with a second person — shared views, assignment, approvals — and offers pre-filled suggestions from the email domain.

Why it works: invitations are the strongest retention signal in collaborative products, but only when the user understands what they gain. Never make this step mandatory; a forced invite is the fastest way to lose a solo evaluator.
Pattern 06 · Recovery The resumable wizard

State is saved at every step, so a user who closes the tab at step two returns to step two rather than the beginning — and a persistent banner offers to resume rather than restart.

Why it works: a meaningful share of abandonment is interruption, not rejection. Resuming costs nothing to build and recovers users your funnel has already written off.

Measuring wizard performance

A wizard is the cleanest funnel in your product: fixed steps, known order, one entry point. Instrument it properly and it will tell you exactly where your onboarding fails.

MetricWhat it tells youWhat to do about it
Step completion rateWhich specific screen loses peopleCut, default, or make that step skippable.
Skip rate per stepWhich questions users consider irrelevantA step skipped by most users should not be in the wizard.
Time per stepWhich question is confusing rather than unwantedLong time plus high completion usually means unclear wording, not a bad question.
Overall completion rateWhether the wizard as a whole is worth its lengthCompare against a shortened variant before adding anything new.
Activation rate by pathWhether completing the wizard actually helpsIf skippers activate at the same rate, the wizard is decoration.

That last row is the one worth taking seriously. Compare activation rates between users who completed the wizard and those who skipped through it. If the two groups activate at similar rates, your wizard is a toll booth rather than a setup step — and the fastest available improvement is deletion.


Building an onboarding wizard with Kompassify

Wizards are usually built into the product itself, which makes them slow to change: every experiment needs an engineering ticket, a release, and a rollback plan. That is why most teams ship one wizard and never iterate on it.

Kompassify lets you build and modify the guidance layer without code, which changes the economics of iteration:

Kompassify is a no-code digital adoption platform for SaaS teams — free up to 100 monthly active users, paid plans from $129/month, GDPR-compliant with EU hosting.

Ship a shorter wizard this week

Move the non-blocking steps into a checklist, branch the experience on one question, and see step-level drop-off — without an engineering ticket.

Start for Free →

Frequently Asked Questions

What is an onboarding wizard?

An onboarding wizard is a short, linear sequence of screens shown immediately after signup that collects the information and choices a product needs before it can be useful, such as a workspace name, a data source, or a first integration. Unlike a checklist or a product tour, a wizard is blocking: the user either completes it, skips it, or abandons signup. That makes it powerful and expensive at the same time.

What is the difference between an onboarding wizard and an onboarding checklist?

A wizard is linear and blocking, runs once at signup, and exists to configure the product. A checklist is non-linear and non-blocking, persists across days or weeks, and exists to drive a sequence of first wins. The most common structural mistake in SaaS onboarding is putting checklist work, such as creating a first project or exploring the dashboard, inside the blocking wizard.

How many steps should an onboarding wizard have?

Two to four steps is the range where completion stays high. Beyond five, drop-off rises sharply, and the later steps perform worst, which is a problem because teams tend to place the most valuable step, such as connecting an integration or inviting the team, at the end. Apply one test to every step: if this answer were missing, would the product be broken, or merely less personalised? Only the first category belongs in the wizard.

Should onboarding wizard steps be skippable?

Yes. A skippable step loses that step, while an unskippable one risks losing the user entirely. Make the skip option visible but secondary, and move any skipped item straight into the onboarding checklist so it is deferred rather than lost. Forcing a team invitation is a particularly reliable way to lose solo evaluators who would otherwise have converted.

Where should an onboarding wizard end?

Inside the product, in a workspace that already has something in it: the first object created, sample data loaded, or the configured view already open. Ending on a summary or congratulations screen, or worse on an empty state, wastes the entire investment, because the user did the work and the product responded with nothing they can act on.

How do you measure whether an onboarding wizard is working?

Track step completion rate, skip rate per step, time per step, overall completion rate, and, most importantly, activation rate split by whether users completed or skipped the wizard. If skippers activate at roughly the same rate as completers, the wizard is a toll booth rather than a setup step, and the fastest improvement available is to remove it.

When should you use a product tour instead of a wizard?

Use a tour when the goal is to explain the interface rather than configure the product. A wizard collects input the product cannot work without; a tour teaches a user what a screen does at the moment they arrive on it. If you find yourself writing explanatory copy inside wizard steps, the content probably belongs in a contextual tour or tooltip instead.

Can you build an onboarding wizard without code?

The blocking setup screens themselves are usually part of the product, but almost everything around them, including the branching, the deferred checklist, the follow-up tour and the step-level analytics, can be built and changed without engineering work using a digital adoption platform. That matters because the main reason teams never improve their wizard is that every experiment requires a release.