This episode is a solo one. I went through two years of onboarding teardowns, from Opal and Finch to Headway, Tiimo, Tolan and Zing, and pulled out the 13 patterns that keep showing up in apps with high-converting flows. Each one comes with real examples of apps doing it well. I've spent over 10 years in mobile growth and monetization, including time at Automattic and Balance, and now run Botsi, where we work on pricing and paywall optimization for subscription apps. The teardowns behind this episode live in Botsi's onboarding flow library.

The job of onboarding: onboarding leads to higher intent, which makes the paywall feel like the obvious next step

Your onboarding has one job: build enough intent that the paywall feels like the obvious next step. It isn't a hallway people walk through on the way to your app. It's where they decide whether your app is worth paying for, so treat every screen like it has a job.

The 13 Principles

1. Your first screen is a drop-off point

A boring default welcome screen next to an interactive hook: Opal's glowing rock that cracks open when you tap it

When teams look at onboarding drop-off, they check three places: sign-up, permissions, and the paywall. Those are big leaks. But the very first screen, the one that loads when the app opens, can lose 10 to 30 percent of new users before they tap anything. Some of those people were never going to stick around. Plenty were, and nothing on the screen gave them a reason to.

Big apps can open on welcome, get started, log in because they have a brand doing the work. Everyone else needs a hook. Opal opens on a glowing rock you tap until it cracks open. Rise Science opens with copy: "Imagine I told you there was a drug that gave you..." and you have to tap to find out it's sleep. Flo plays a short animation before asking for any of its sensitive health data.

If you need a login button, keep it, but follow Finch. The main button for new users says "Hatch a new pet," and the next screen has you pick an egg color. Make it interesting, and make it easy.

2. Borrow UI patterns people already know

Headway's intro screens rebuilt to look like Instagram stories, which you tap to advance or let autoplay

You don't need to invent a new interface for your intro screens. Headway switched its first screens to look like Instagram stories: you tap to advance, or let them autoplay. Zing made the same change for the same reason, because those first screens have the worst drop-off in the flow.

Think about why the stories pattern spread in the first place. It's simple and it's engaging, and social media apps are the kings of sticky engagement patterns. People already know how it works. Zing also made its opener friendlier at the same time, rewriting "Zing Coach" to "Meet Zing, your AI coach." Small edit, much cleaner.

Every pattern a user learned in another app is one less thing they have to figure out in yours.

3. Ask for the account when intent is highest

Tiimo's account creation in 2024 sitting right before the paywall, compared with 2025 where it moved up front after the welcome screen

Tiimo won Apple's App of the Year in 2025, and one of the changes it made was moving account creation. In 2024 it sat right before the paywall. Now it comes right after the welcome screen. According to their team, account creation and trial starts both went up.

Two things are happening. Intent is highest at the start of the flow, and if someone won't create an account then, they're unlikely to pay later. Moving the ask also separates the two biggest asks, so users never hit an email form and a payment screen back to back.

This one deserves a test, not a blind copy. Either ask early, while intent is high and the paywall is far away, or skip it and only ask after the paywall. If you do ask, offer Sign in with Apple, Google, and email. SSO usually wins, but not always.

4. Say the user's goal back to them

Asking the user's goal, then repeating their answer back as the value proposition on a later screen

Lots of apps ask "What's your goal?" and never mention the answer again. That's a waste of the most useful thing a user tells you.

Ask the goal, then ask a follow-up about why that goal matters to them. Then show a screen explaining how the app delivers on that exact answer. Zing repeats "muscle gain" throughout the flow if that's what you picked. Tiimo, a productivity app for neurodivergent people, changes the paywall headline if you say you're neurodivergent. Impulse sells fitness or savings on the paywall depending on your goal. Use their name too. Everyone likes reading their own name.

Complex personalization is hard to get right. Putting a user's own words back in front of them almost always works, and I've rarely seen it fail to lift conversion. The same logic carries further down the funnel, into how you word the paywall for different segments.

5. Make the questions feel like a conversation

Opal's onboarding quiz rebuilt as a chat, with questions arriving as messages in a conversation

Your quiz doesn't have to feel like a stale, stuffy form. Opal rebuilt its whole quiz as a chat experience. It isn't a real chatbot, and the questions are the usual ones: your name, what matters to you, your daily screen time. But because they arrive as messages in a conversation, you actually read them instead of going tap, tap, tap and ignoring everything just to get through it.

Tolan goes further with a voice experience where you answer out loud, and it converts very strongly because it takes the conversational idea to its limit. Zing lets you pick a coach, and that coach then walks you through the rest of the onboarding like a mascot guiding you.

Don't go overboard and add tons of screens. Be friendly, be conversational, and build the experience like two people talking.

6. Onboarding should feel like the product

Finch's onboarding where you design a bird by picking its color, name and personality

People value what they put work into. So give them something to build. Finch has you pick your bird's color, gender, name, and personality. None of it is about you, which means no privacy concerns and no awkward questions, but you still end up attached to a bird you made.

Rosebud, an AI journaling app, lets you write a full journal entry before the paywall. You're getting value from the product before you're asked to pay for it. That's the goal: onboarding as an extension of the product, not a survey about the product.

One practical detail. If you ask users to type something, give them a way out of the blank field. Finch puts a shuffle button next to the name input so nobody stalls trying to think of a name.

7. Show them the math

Opal turning your screen time and age into a single line: Opal will get you back five years of your life

Opal takes your daily screen time and your age and turns them into one line: "Jacob, Opal will get you back five years of your life." That's the aha moment, and it happens before the paywall.

Rise Science pulls data from your phone and Apple Health to calculate your sleep debt, so the number is grounded in real data. Headway asks how much time you'll spend per day and tells you 15 minutes a day becomes 38 book summaries a month.

Find a way to express the value of your app as a number the user will actually receive. When they reach the paywall, they should already know what they're getting and how much of it.

8. Build trust across the whole flow

Flo answering its users' biggest fear by returning to data privacy across its onboarding screens

This may be the most important item on the list. Most apps cram trust-building into the paywall. The good ones build it from the first screen on. Headway opens with "#1 most downloaded" next to its benefits and shows social proof with your name on it right after you create an account.

Find your users' biggest fear and address it, more than once. For Flo, that fear is health data privacy, so its loading screens keep cycling back to privacy. Lead with emotion, then back it up with a number. People decide with emotion and justify with logic. Endel puts a single stat before the paywall: 71 percent of users report better sleep.

Three kinds of social proof: a large user count for scale, one specific testimonial for a result, and press logos

Mix up your social proof: a big user count (JustFit flashes 1M+ users and a 4.8 rating), one specific testimonial, press logos. If you have a strong proof point like App of the Day, repeat it. Some users will miss it the first time.

9. Earn the permission opt-in

Permission prompts are a major drop-off point. Opal builds its own screen before the system prompt, with arrows showing exactly where to tap. If you need several permissions, copy that.

If you only need push, consider skipping the primer. Fire the native prompt and most people will tap yes. Zing takes a middle path and shows you what the notification will look like before asking.

Headway has iterated on this more than any app I've seen. In 2024 it had a preference center where you picked notification types. In 2025 that became one question: what time of day do you want to read? The prompt fires right after. Timing is worth testing too. Imprint asks right after sign-up, Flo explains the value first, and JustFit builds excitement for a few screens before asking.

10. Use video, because it's cheap now

Video used to mean a studio, and ten or twenty thousand dollars in editing and reshoots. AI has made animation and video production cheap enough that there's no real excuse left, and you can make something that still feels authentic.

Lingokids runs an animated video inside its paywall to cycle through value props. BoldVoice puts a real instructor on camera to walk you through the onboarding questions, which helps you understand what the product actually does.

Ladder has the most polished video of any app I've seen, and it positions them as premium while attaching you to their coaches. Dr. Kegel is the extreme case: almost 90 percent of its onboarding is one testimonial video that feels like an old TV infomercial. Tolan's animated intro is worth downloading the app just to watch.

Motion and movement create a perception of quality that static screens don't. It isn't a new idea, but a lot of teams still skip it.

11. Turn loading screens into selling screens

A loading screen split into several progress bars, each naming a different part of the plan being built for the user

A lot of loading screens are fake. Apps load fast now. Either way, the screen is there, so use it.

My recipe: title it around personalizing their plan. Split it into multiple progress bars, one per piece of the app being built for them. Rotate through the values you talked about earlier in the flow, and sprinkle in testimonials, ratings, and press mentions. Flo cycles through its privacy values; JustFit goes back to its user counts and ratings.

Then add one more step before the paywall. JustFit and Zing show a plan preview right after loading, so users know exactly what they're getting. Headway uses a hold-to-commit screen. Each one deepens commitment right before you ask for money.

12. Make the trial impossible to miss

Tolan's multi-page paywall spreading the trial timeline across several screens before showing the plans

People don't read. They tap through, hit the paywall, think "I'm being charged," and leave, even when the free trial is clearly visible. A lot of paywall tactics come down to one job: making sure the user notices the trial.

PictureThis makes users choose between trial options, which guarantees they see that a trial exists while steering them toward annual. Opal lets you pick the day you get your trial reminder, which forces acknowledgment and reduces people canceling right after they start. Tolan and Headway use Duolingo's multi-page paywall, spreading the trial timeline across pages with the reminder language in big text.

Most people will still close the paywall. Don't give up. Follow with a second offer: a longer trial, a discount, a monthly option like Headway, or a cheaper Lite plan like Tolan.

13. Onboarding doesn't end at the paywall

Imprint skipping the home screen and dropping a new subscriber straight into their first lesson

Someone just paid you. Don't drop them on an empty or confusing home screen. The goal now is to stop them from canceling the moment their trial starts.

Imprint skips the home screen and puts you straight into your first lesson. Tiimo replaced its old setup with five focus steps. Finch pre-fills a few easy goals so you hit the aha moment immediately and have a reason to come back tomorrow. Balance asks people who picked sleep if they're ready to sleep now, then auto-starts the meditation. Tolan pays you tokens for scheduling a check-in.

Onboarding continues after payment: trial start leads to a first win, which leads to a return, which leads to retention

It's more work to build, but this is what moves you from a good trial start rate to a high trial conversion rate and strong activation. Retention starts here.

Key Takeaway

Don't try to ship all 13 next week. Pick the one you think will make the biggest difference, test it, and repeat. If you want somewhere to start, we keep a running list of A/B tests for your onboarding flow. Track onboarding completion, the share of users reaching the paywall, trial conversion, and day 7 and day 30 retention. A longer flow is fine if it's a better one.

The idea worth keeping from all of this: your onboarding isn't a hallway people walk through on the way to your app. It's where they decide whether your app is worth paying for. Treat every single screen like it has a job.