Digital accessibility: what the law really requires
4 min read
Key takeaways
- The first screen must show the value, not ask for it: postpone sign-up as long as possible.
- Each additional introduction screen costs about 20% of the remaining users.
- Ask for a permission at the moment its usefulness is obvious, never at launch.
- A well-designed empty state beats a tutorial: show an example, not an explanation.
- Measure the time to the first successful action: it is the activation metric that counts.
On most of the apps we audit, between 40 and 60% of the people who install never come back after the first launch. The product is not at fault: they did not stay long enough to judge it.
Onboarding is not window dressing. It is the moment where retention is decided, and it plays out in less than sixty seconds.
The rule that overrides all the others
Show the value before asking for anything.
The default order is almost always wrong: welcome screen, three explanation screens, account creation, permission requests, and finally the app. At every step, some of the users leave, and they leave before having seen what you can do.
The order that works: the app, straight away, with sample data if necessary. Then, when the user wants to keep, share or personalise something, and only then, account creation.
The order of magnitude to remember: each screen added before the first useful action costs about 20% of the remaining users. Three introduction screens means half of your audience lost before they have seen anything.
The six critical moments
1. The very first screen
It answers a single question: what can I do here? Not "who are we", not "our values", not a carousel of benefits.
The best answer is often the app itself, pre-filled with a realistic example. A user understands in three seconds what no text can explain in three screens.
2. Account creation
Push it back as far as possible. When it becomes necessary, explain in one sentence why, at that precise moment: "Create an account to find your favourites on all your devices."
Reduce it to the bare minimum: email and password, or a magic-link login. Every extra field costs users, and name, address or company can be asked later, when they actually serve a purpose.
3. Permissions
Notifications, location, camera, contacts. Requested at launch, they are overwhelmingly refused, and a refusal is almost final, since the user then has to go through the system settings.
Ask for them at the moment their usefulness is obvious: location when the user is looking for a nearby point of sale, notifications just after a booking. The acceptance rate changes completely.
4. The empty state
This is the most important screen in your app, and the one that is designed last, in a rush.
A good empty state does not say "no items". It shows what the screen will look like once filled, and offers the action that gets you there. It is a tutorial that does not present itself as one, and one of the components that benefit from being standardised in a design system.
5. The first successful action
Identify the action that proves the value of your product: the first booking, the first document imported, the first message sent. The whole design of the onboarding must aim to bring the user to it as fast as possible, removing everything that gets in the way.
It is also your activation metric: what proportion of new users reach it, and how long it takes.
6. The return, the next day
Onboarding does not stop at first launch. The second session is where the habit is formed or not. A concrete reason to come back (a useful reminder, a result that is ready, a new piece of information) is worth more than ten explanation screens on day one.
40–60%Of users lost after a single launch, without a carefully crafted onboarding
≈ 20%Of users lost for each screen added before the first action
< 60 sThe time you have to prove the value
What to measure
Three metrics are enough, and they can be set up in a day.
The activation rate. The proportion of new users who complete the first useful action. This is the figure to improve first.
The time to activation. Median, not average. If it exceeds two minutes, the journey contains one step too many.
Drop-off by screen. On which exact screen do users leave? A screen that loses more than 25% of its audience is a screen to rework, not to document.
The most costly mistakes
The introduction carousel. Three to five screens explaining the benefits of the product. Overwhelmingly skipped, and those who read them do not remember them.
Asking for a rating too early. Asking for a rating before the user has gained value produces exactly the opposite of the intended effect.
The full profile form. First name, last name, phone, company, sector, size, before the very first use. Every field costs you users, and the information can almost always be collected later.
The silent loading screen. Three seconds with no indication is a broken app from the user's point of view. A loading state that shows the structure to come changes the perception of the waiting time.
Nobody opens an app to discover an app. People open it to do something. Your onboarding succeeds when it disappears behind that something.
Where to start
Install your own app on a clean phone, stopwatch in hand. Count the number of gestures and the time it takes to complete the first useful action.
Then have three people who do not know the product do it, watching them without helping. In thirty minutes you will know exactly what to fix, and it will rarely be what you thought.
Sources
- Onboarding — Apple (Human Interface Guidelines)
- Privacy — Apple (Human Interface Guidelines)
- Demander des autorisations d'exécution (in French) — Android Developers
