---
title: "Onboarding That Sticks: The First Minutes in Your Product"
url: "https://insideuxdesign.com/en/blog/onboarding-that-sticks-the-first-minutes-in-your-product"
language: "en"
description: "Why product tours rarely help and how new users reach their first success faster: 7 onboarding levers with a worked example. Start improving yours today."
published: "2025-08-06T09:00:00+02:00"
updated: "2026-10-15T00:38:56+02:00"
---

# Onboarding That Sticks: The First Minutes in Your Product

New users don't want to click through tutorials. They want to get something done. Learn how to define the first success in your product and shorten the path to it with seven levers, explained through one running example.

Picture a freelancer trying out a new invoicing tool one evening. They have exactly one goal: send the invoice for last week's project. After signing up, five welcome slides with illustrations appear. They swipe them away. Next comes a questionnaire about industry, team size and “How did you hear about us?” Then an empty dashboard full of charts without data, covered by a product tour with nine speech bubbles. At bubble four, they click “Skip.” Ten minutes later they still haven't created a single invoice, but they have opened two other tabs. One of them is a competitor's website.

This example is made up, but the pattern is common. Many products put a lot of work into onboarding and still lose people at exactly this point, right after those people deliberately chose to try them. The problem is rarely too little explanation. Usually it's too much explanation in the wrong place and too little of a path to the first success.

This article shows why classic tutorials often achieve little, how to define the first success in your product, and which levers help new users get there faster. The invoicing tool serves as a running example throughout.

## Why the first minutes matter so much

[Onboarding](https://insideuxdesign.com/en/glossar/onboarding) covers everything that helps people understand a product and get their first value from it: from sign-up through setup to the first completed task. So it isn't a single screen, it's a journey.

Along that journey, people behave differently than many teams would like. John M. Carroll and Mary Beth Rosson described this back in 1987 as the “paradox of the active user.” Their observation from studies at IBM: people don't read manuals, they start right away. They want to get their current task done and skip introductions, even when those would save them time in the long run. Jakob Nielsen later picked up the paradox at the Nielsen Norman Group (NN/g) and drew a clear conclusion from it: design for the way people actually behave, not for an idealized rational user.

For onboarding, that means the people opening your product right now aren't in learning mode. They're in getting-things-done mode. Every slide, every questionnaire and every speech bubble stands between them and the reason they came. At the same time, they don't yet know whether the product is worth it. Their patience is especially short at this moment, because nothing is holding them yet.

This leads to the goal every good onboarding has to be measured against: new users should reach a real, tangible success as quickly as possible. Not “know their way around the interface,” but have actually accomplished something that's useful to them.

## What tutorials can and can't do

Many teams instinctively start with an introduction: a carousel of slides, a video, or a tour that points at one control after another. It feels considerate. The research, however, is sobering.

In 2020, NN/g ran a quantitative study with 70 participants across four iOS apps. One group always saw the app's tutorial, the other always skipped it. Then both groups completed the same tasks. The result: success rates were practically identical at 91 percent (with tutorial) and 94 percent (without), and the difference wasn't statistically significant. Task time showed no significant difference either. Only one thing was significant: the tutorial group rated the tasks as harder. The author, Alita Kendrick, concludes that tutorials made people neither faster nor more successful and advises thinking twice before building a tutorial for simple apps.

That's not because explanations are useless in principle. It's about timing. Someone reading a slide about importing customer data doesn't have that task in front of them. They have to keep the information in their head until they need it, and by then it's usually gone. That puts a load on working memory without achieving anything.

NN/g therefore distinguishes between two kinds of proactive help. “Push revelations” appear regardless of what someone is doing, such as an overlay at launch. They're often ignored because they interrupt the actual task. “Pull revelations” appear in a relevant context instead, for example a hint that shows up when someone opens a feature for the first time. They're less likely to be ignored because they help exactly when the question comes up. NN/g's recommendation is clear: favor pull over push.

That doesn't mean tutorials are forbidden. For complex tools or unfamiliar interaction concepts, a short introduction can make sense. But it's the last resort, not the first. If a product has to be explained before anyone can use it, the problem usually lies in the interface, and that's where the work should start.

## Defining the first success

### What is the first success in your product?

Before you tinker with onboarding, your team needs a shared answer to one question: which moment shows a new person that the product is worth it for them? Product teams often call this the “aha moment” or “activation,” and the time it takes to get there “time to value.”

A good first success meets three conditions:

- **It delivers real value** to the person, not just to your metrics. “Profile 100 percent complete” isn't a success from the user's point of view.
- **It's reachable in the first session**, ideally within a few minutes.
- **It's a preview of regular use.** Anyone who has experienced it once understands why it's worth coming back.

For the invoicing tool, the answer is obvious: the first invoice has been sent, or at least downloaded as a finished PDF. For a project management tool, it might be the first task assigned to another person. For a learning app, the first completed lesson. For an analytics tool, the first chart built from your own data.

It helps to think in terms of the job people want to get done with your product. Someone opening an invoicing tool doesn't want to “manage invoices.” They want to get paid for work they've done, with as little effort as possible. The first success is the smallest step that visibly moves that job forward.

### Making the path visible

Once the first success is defined, a simple but often uncomfortable exercise follows: walk through your own product as a new user and count every step until that success. Every click, every field, every slide, every wait counts.

In the original version of the invoicing tool, the path looks like this:

1. Sign up with email, password and password confirmation
2. Confirm the email address (switch to the inbox)
3. Five welcome slides
4. Questionnaire with four questions about industry, team size, goal and referral source
5. Fill in all company details: name, address, tax number, bank details, logo
6. Product tour with nine speech bubbles
7. Empty dashboard, with “New invoice” hidden in the menu
8. Create a client first, only then is an invoice possible
9. Create and send the invoice

Nine stages, of which only the last two contribute directly to the goal. On top of that, the email confirmation forces a switch to another app, and many people never come back from it. This list is your most important tool for the redesign. For each step, ask the same question: does it have to happen before the first success, or can it come later, be optional, or disappear entirely?

## Seven levers for onboarding that sticks

### 1. Move sign-up further back

Many products require an account before they show anything at all. Yet the barrier is never higher than at the very beginning, when nobody knows whether the product is worth it. The alternative is known as “gradual engagement” or deferred sign-up: people try things out first, and the account is created once there's something worth saving.

A well-documented example comes from Duolingo. Gina Gotthilf, who led growth at the language learning app at the time, described an A/B test in an interview with First Round Review: new users could complete a lesson before being asked to sign up. According to Gotthilf, simply moving the sign-up screen back a few steps led to roughly a 20 percent increase in daily active users.

Not every product can start without an account, for instance when data has to be stored securely from the start. But even then, you can lower the barrier. In the invoicing tool, an email address or signing in with an existing account is enough to get started. Email confirmation is no longer forced up front, only when the first invoice is actually about to be sent by email. Until then, the person can work in peace.

### 2. Only ask what improves the first success

Up-front questionnaires are popular because they deliver data for marketing and sales. For the people filling them in, they're work with no visible return. A question earns its place in onboarding only if the answer noticeably improves the next few minutes.

In the invoicing tool, the questionnaire is therefore almost entirely gone. A single question remains: “Do you charge VAT on your invoices?” The answer genuinely determines what the first invoice looks like. In Germany, for example, small businesses can opt out of charging VAT under the so-called small business scheme (Kleinunternehmerregelung), and their invoices then need a note saying so. “How did you hear about us?” moves to a later, optional email.

The same goes for company details. Instead of requiring everything up front, the tool asks for information where it's needed: name and address when creating the first invoice, bank details in the invoice footer, the logo as an optional step afterwards. How to design the remaining fields so nobody gives up is covered in detail in [Forms Nobody Abandons](https://insideuxdesign.com/en/blog/forms-nobody-abandons-12-rules-for-better-form-ux).

Permissions on smartphones, such as notifications, camera or location, are a special case. In its Human Interface Guidelines, Apple recommends requesting them during onboarding only if the app can't function meaningfully without them, and otherwise waiting until the situation actually requires the permission. In the invoicing tool, a notification request only makes sense once the first invoice has been sent: “Want us to let you know as soon as the invoice is paid?” At that moment, the benefit is obvious.

### 3. Learning by doing instead of by slides

If an introduction is necessary, it works best as an action. Apple puts it concisely in the Human Interface Guidelines: people learn better by performing tasks than by reading instructions. Interactive introductions also get them using the app right away.

In the invoicing tool, a guided first run replaces the five slides and the nine-part tour. After sign-up, the person doesn't land on the dashboard but directly in the invoice editor, with a short heading: “Create your first invoice. Takes about two minutes.” The fields are arranged in the order they're needed. The client can be created right inside the editor instead of requiring a detour through client management first.

Two rules apply to any kind of introduction:

- **Skipping must always be possible.** Anyone who just wants a quick look shouldn't be trapped in a flow. Apple also recommends not showing a skipped tutorial again on every launch, but keeping it available in a findable place such as help or settings.
- **Needing no explanation beats a good explanation.** If a step only makes sense with a speech bubble, first check whether a clearer label, a better button or a more sensible order would make the explanation unnecessary.

### 4. Use empty states as a launch pad

A new account is empty: no invoices, no clients, no revenue. This situation is called an [empty state](https://insideuxdesign.com/en/glossar/empty-state). In many products it's an afterthought, like a table without rows or a chart with a flat zero line. Yet it's the place new users pass through most often.

Kate Kaplan of NN/g describes three guidelines for empty states in complex applications: they should communicate system status (is it still loading, did an error occur, or is there simply no data yet?), they should provide learning cues, and they should offer direct pathways to key tasks. Kaplan considers this approach more successful in most cases than forced tutorials at first launch.

Here's how that looks in the invoicing tool:

> **Before:** An empty table with the column headers “No.,” “Client,” “Amount,” “Status” and below it the text “No entries found.”
> 
> **After:** “Your invoices will appear here once you've sent the first one. You'll see at a glance what's been paid and what's still outstanding.” Below it, a prominent “Create first invoice” button and a smaller “Import existing invoices” link.

The new text explains what the area is for, and the button leads straight to the first success. That's good [microcopy](https://insideuxdesign.com/en/glossar/microcopy): short, concrete, task-focused. The same applies to the dashboard. Instead of empty charts, it initially shows a single area with the next sensible step.

### 5. Start with templates and sample data

For many people, a blank page is harder than a half-finished one. Templates and sample data lower the hurdle of the first step and show what the result can look like.

That's why the invoicing tool's editor gets a pre-filled invoice template with sensible defaults: today's date, a sequential invoice number, payment terms of 14 days and a first line item with the placeholder “Service, e.g. web design for homepage.” The person only has to replace, not invent.

With sample data, one boundary matters: it has to be clearly recognizable as an example and removable with one click. A sample invoice that accidentally goes to a real address or shows up in the revenue statistics destroys exactly the trust onboarding is supposed to build. A clear label like “Example, will not be sent” and a “Remove examples” button solve this.

### 6. Contextual help instead of a tour in advance

The invoicing tool can do more than create a first invoice: recurring invoices, payment reminders, quotes, exports for the tax accountant. Nobody needs to know all of that in the first minute. This is where progressive disclosure helps, meaning information and options appear only once they become relevant.

Instead of a tour that points at everything in advance, small hints appear at the right moment. When the third invoice goes to the same client, a subtle hint appears: “Do you bill this every month? A recurring invoice takes care of it automatically.” When an invoice passes its due date, the tool suggests a payment reminder. These are pull revelations in NN/g's sense: help that responds to a real situation.

A few rules of craft apply to such hints:

- **One hint at a time.** Three overlapping tooltips aren't help, they're noise.
- **Dismissed means dismissed.** A hint that comes back on every visit after being closed becomes an annoyance.
- **Make it accessible.** Under WCAG success criterion 1.4.13 (“Content on Hover or Focus,” level AA), content that appears on hover or focus must be dismissible without moving the pointer, for example with the Escape key. It must not disappear when the pointer moves onto it, and it must stay visible until the person dismisses it or leaves the trigger. Hints that fade out on their own after a few seconds disappear too quickly for many people.

### 7. Show progress, but honestly

Some products need more than one step before they become truly useful. A short checklist that shows what's done and what would make sense next can help. In the invoicing tool, it only appears after the first success and contains just three items: “First invoice created” (already checked), “Add your logo” and “Connect your bank account for automatic payment matching.”

The fact that the first item is already checked has a psychological basis. In 2006, Ran Kivetz, Oleg Urminsky and Yuhuang Zheng studied the so-called goal-gradient effect in the Journal of Marketing Research: people put in more effort the closer they get to a goal. In a field experiment with a café's reward cards, one group received a card with ten slots, the other a card with twelve slots, two of which were already stamped. Both had to buy ten coffees. The group with the apparent head start reached the goal faster.

In onboarding, this works as long as the progress is real. The first checkmark in the invoicing tool stands for a task that was actually completed. A progress bar that artificially starts at 30 percent, or a checklist that suddenly adds new items after the last one, quickly tips over into manipulation. The rule of thumb: progress indicators should tell the truth, just in a friendly way.

## Onboarding doesn't end after the first session

The first success is a beginning, not an end. The next critical point is the second session: does the person come back, and do they find their way back to what they achieved the first time?

Three things help. First, a start that picks up where the last session left off. In the invoicing tool, the dashboard shows the sent invoice and its status on the second visit (“Opened on Tuesday, not yet paid”) instead of a generic welcome. Second, messages outside the product that have a real reason. An email saying “Your invoice has been paid” brings people back because it's good news, not because it's meant to remind them of the product. Third, the ability to find help again later. Anyone who skipped the introduction should be able to call it up again from help or settings.

Existing customers need onboarding too: when a new feature launches, the same rules apply. A hint in the right context works better than a login pop-up that lists every new feature at once.

## How to check whether your onboarding works

Whether onboarding is good can't be decided in a team meeting. People who use a product every day can hardly imagine what it feels like the first time. Two methods complement each other well.

**Usability tests with people who have never seen the product.** Give them a realistic task that matches the first success, such as “Last week you built a website for a café. Send them the invoice.” Then watch without helping. Where do they hesitate? What do they skip? At what point do they ask what comes next? Even a handful of sessions usually reveals the biggest hurdles.

**Metrics along the journey.** Measure how many new accounts reach the first success (often called the activation rate), how long they take on average, and at which step most people drop out. These numbers don't show why something goes wrong, but they show where to look. And they make it visible whether a change actually improved anything.

A note on interpretation: if many people skip a tour, that's not the people's fault, it's information about the tour. And if a step has a high drop-off rate, the first question isn't “How do we explain it better?” but “Do we need it here at all?”

## Conclusion

In the redesigned invoicing tool, the freelancer from our example signs up with their email address and lands directly in the editor. A single question about VAT adapts the template, the fields are pre-filled, and the client is created right inside the invoice. A few minutes later, the first PDF is ready. Only now do they confirm their email address, because they want to send the invoice. A short checklist suggests sensible next steps, and the remaining features show up when they're needed.

Nine stages have essentially become two: sign up and create the first invoice. None of this is spectacular, and that's exactly the point. Good onboarding doesn't feel like onboarding. It feels like the product understood why you came.

If you take away just one thing from this article, make it this: define the first success in your product, count the steps to get there, and cut everything that doesn't absolutely have to happen before it.

## Checklist

- The team has clearly defined the first success, from the user's perspective and reachable in the first session.
- Every step up to the first success is listed, and each one is truly necessary.
- Sign-up happens as late and is as short as possible, and email confirmations don't block getting started.
- Onboarding questions directly improve the next few minutes, everything else comes later.
- Permissions are only requested once their benefit is obvious.
- Introductions are interactive, short and skippable at any time, and can be found again later.
- Empty states explain what an area is for and lead to the next task with a button.
- Templates and sample data make getting started easier and are clearly labeled as examples.
- Hints appear in the right context, one at a time, and are accessible.
- Progress indicators and checklists are honest.
- New people have gone through the onboarding in a usability test, and activation rate and drop-off per step are measured.

## Sources

- [Nielsen Norman Group: Mobile Tutorials: Wasted Effort or Efficiency Boost?](https://www.nngroup.com/articles/mobile-tutorials/)
- [Nielsen Norman Group: Paradox of the Active User](https://www.nngroup.com/articles/paradox-of-the-active-user/)
- [Nielsen Norman Group: Help and Documentation (Usability Heuristic #10)](https://www.nngroup.com/articles/help-and-documentation/)
- [Nielsen Norman Group: Designing Empty States in Complex Applications: 3 Guidelines](https://www.nngroup.com/articles/empty-state-interface-design/)
- [Apple Human Interface Guidelines: Onboarding](https://developer.apple.com/design/human-interface-guidelines/onboarding)
- [First Round Review: The Tenets of A/B Testing from Duolingo’s Master Growth Hacker](https://review.firstround.com/the-tenets-of-a-b-testing-from-duolingos-master-growth-hacker/)
- [Kivetz, Urminsky, Zheng (2006): The Goal-Gradient Hypothesis Resurrected, Journal of Marketing Research](https://business.columbia.edu/sites/default/files-efs/pubfiles/1200/goalgradient.pdf)
- [W3C: Understanding Success Criterion 1.4.13 Content on Hover or Focus](https://www.w3.org/WAI/WCAG22/Understanding/content-on-hover-or-focus)