App Onboarding: Helping New Users Reach Their First Useful Action
Good onboarding is mostly about what you leave out. A new user installed your app to do something specific, and onboarding's job is to get them there as quickly as possible – then teach the rest in context, when it is needed. Research by the Nielsen Norman Group found that the swipe-through tutorials many apps show at first launch didn't help people complete tasks, and Apple's design guidelines recommend teaching through interaction and postponing non-essential setup. In practice: define the first useful action, remove every step that doesn't lead to it, ask for permissions and accounts only when they are needed, and measure how many new users get there.
Define the first useful action
The first useful action is the moment a new user gets something of value – not the moment they finish registering. Some examples:
| Type of app | First useful action | What can wait |
|---|---|---|
| Shop or ordering | Find a product and see its price and availability | Account, saved addresses, loyalty programme |
| Bookings (salon, clinic, service) | See free slots and book one | Profile details, notification preferences |
| Catalogue or reference data | Search and open the right item | Favourites, sync between devices |
| Measurement or tracking | Complete a first measurement | Rankings, sharing, history charts |
| Customer account (utility, insurance) | See the current balance or status | Document archive, settings |
Write yours down and count the screens and taps between opening the app and reaching it. That number is your starting point. If the action is something customers would do only once, a good mobile website may serve them better than an app – see Does Your Business Need an App or a Better Mobile Website?.
What usually stands in the way
Swipe-through tutorials
In a study with 70 participants and four iPhone apps, NN/g compared people who read the "deck of cards" tutorial at first launch with people who skipped it. Task success was comparable (91% versus 94%), the tutorials didn't make people faster, and those who read them rated the tasks as harder. NN/g's advice is to avoid creating onboarding whenever possible and invest in a usable interface instead: "Instructional onboarding should not be used to supplement poor design." If you keep a tutorial, make it short and easy to skip. Apple adds that a skipped tutorial shouldn't reappear on later launches but should stay easy to find, for example in a help or settings area.
Login walls
Asking for registration before the user has seen any value is one of the most expensive steps you can add. Apple's App Review Guidelines say: "If your app doesn't include significant account-based features, let people use it without a login." Even where accounts are needed, let people look around first and ask for an account at the moment it is required – to place an order, save a favourite or see their own data.
Permission requests at launch
A series of system dialogs at first launch – location, notifications, camera – asks for trust before the app has earned it. Android's guidance is to "ask for a permission in context, when the user starts to interact with the feature that requires it", never to block the user, and to keep the app usable if permission is refused. There is a practical reason too: on Android, if the user taps Deny for the same permission more than once, the system stops showing the dialog, and the app loses the chance to ask again. Apple's guidance is similar: request permission during onboarding only if the app can't work without it; otherwise ask when people first use the feature that needs it.
For notifications, new installs on Android 13 and later have them switched off until the user grants permission, and Google recommends letting people get familiar with the app first – asking, for example, when they tap a bell icon or submit an order.
When permission is refused, offer an alternative. Apple's guidelines give the example of letting people enter an address manually if they decline to share their location.
Setup screens
Choosing themes, interests and preferences before using anything delays the first useful action. Apple recommends postponing "nonessential setup flows or customization steps" and providing reasonable defaults. NN/g notes that content customisation can belong at the start when it makes the experience relevant – choosing a city in a local services app, for example – but visual customisation doesn't, and you should explain why you are asking.
What to do instead
- Open on something useful. The first screen should already let users start the core task – a search box, a product list, today's free slots.
- Teach in context. A short hint the first time someone reaches a non-obvious feature beats a tour of everything. Apple suggests "a collection of context-specific tips instead of a single onboarding flow".
- Use empty states. An empty favourites list or history screen is the natural place to explain what will appear there and how.
- Let people try safely. Sample data or a practice run lets them learn by doing – "teach through interactivity", in Apple's words.
- Ask for ratings and purchases later. Apple advises letting people experience the app before prompting for either.
- Keep legal text out of the flow. Apple recommends letting the App Store display agreements and disclaimers rather than putting licensing details into onboarding.
A worked example
Take a measurement app like roadmeter.app, which records trips, 0–100 km/h acceleration, braking and track laps using the phone's GPS and sensors. Its first useful action is a completed measurement. Location access is genuinely needed for that, so the natural moment to explain and request it is when the user chooses to start a measurement – not on a welcome screen. Rankings, history and sharing only make sense once there is a result, so they can be introduced after the first measurement is saved. The same reasoning works for any app: find the first result, and move everything else to the moment it becomes relevant.
Measure whether it works
Track a simple funnel for new users:
- first app open;
- each onboarding step, if there are any;
- permission prompts shown and accepted;
- account created, if required;
- first useful action completed;
- a return visit within the following week.
Find the step where most people drop out and fix that first. Then watch a few people install and use the app for the first time without help – where they hesitate tells you more than any dashboard.
Checklist
- The first useful action is defined, and the path to it is as short as possible.
- No tutorial – or a short, skippable one that can be found again later.
- No login before the user has seen value, unless the app is built around accounts.
- Permissions requested in context, with a clear reason and a fallback if refused.
- Sensible defaults instead of setup screens.
- Hints and empty states explain features where they are used.
- Rating and purchase prompts only after real use.
- A funnel that shows where new users stop.
If your app loses people between installing it and doing anything useful, the cause can often be found in the first few screens. We can review the first-run experience of an existing app, or design it into a new one, as part of our mobile app development work. Contact us and tell us what your users should achieve first.
Sources
- Mobile-App Onboarding: An Analysis of Components and Techniques – NN/g
- Mobile Tutorials: Wasted Effort or Efficiency Boost? – NN/g
- Onboarding – Human Interface Guidelines – Apple Developer
- App Review Guidelines – Apple Developer
- Request runtime permissions – Android Developers
- Notification runtime permission – Android Developers