An onboarding people finish: lessons from IMPERIA THE1
People finish an onboarding when every screen gives them something back: a reply in a real voice, a choice that changes what comes next, and a clear first action at the end.
People finish an onboarding when every screen gives them something back: a reply, a choice that changes what comes next, or a step that feels like the product has already started. In IMPERIA THE1, the fitness app I built for the brand of Aram, a coach and the face of IMPERIA THE1, the onboarding is an interactive film with branching. It mixes video clips, questions answered with buttons and a step where the user gives their name.
This article covers the design decisions behind that format and how you can apply them to an app for a clinic, a gym, a rental company or a venue in the UAE. I don't publish completion rates for Imperia here, so treat this as a description of the mechanics. Results are a question to put to any developer, me included. The full build of the app, the site and the Apple Watch screens is in the Imperia case study.
Decide what the onboarding must produce
Before anyone designs a screen, write down two things. The first is what the app needs to know to make the first session useful. The second is the one action the user should take straight after: start a workout, book a slot, add a card.
Then map every question to what it changes:
| App | Question | What it changes next |
|---|---|---|
| Fitness app, using IMPERIA THE1's settings | How many days a week can you train? | The flexible week, set from 2 to 6 training days |
| Fitness app, using IMPERIA THE1's programs | What kind of training do you want? | Which program to suggest: THE 4×4, THE RUN, THE HOME, or a "custom workout" with no plan |
| Hypothetical example: car rental | Do you hold a UAE licence? | Which documents the booking screen asks for |
| Hypothetical example: clinic | Is this your first visit? | Which clip plays and which form opens |
Any question without an entry in the third column belongs in settings, not in onboarding.
Why Imperia's onboarding is a film
IMPERIA THE1 is built around a coach. Aram is the brand, and the Pro plan includes a chat with a mentor. A carousel of feature screens would list what the app does. A film lets the coach speak to the user from the first second.
The onboarding plays as a sequence of short clips. The coach speaks, then the screen asks a question with button answers, and the next clip depends on the button. I shot and edited the film myself. That kept the script, the takes and the code with one person, so nothing was lost in a handover between a video crew and a developer.
The format suits any business where trust sits with a person. Hypothetical example: a dental clinic in Dubai opens its app with the lead dentist explaining the first visit in 20 seconds. The app then asks "Is this your first visit with us?" and branches into two different clips.
Branching: every answer moves the story
A branching onboarding works like a conversation. The user answers, the coach responds to that answer, and the user sees that the answer mattered. Each tap produces something new, and that keeps people moving.
Rules for planning branches:
- Keep buttons to two to four options. With more, the user starts reading instead of answering.
- Write the reply clip for every option before filming. If you can't write a distinct reply, merge the options.
- Let branches rejoin. A tree that never merges multiplies filming and editing time with every question.
- Tie each branch to a real setting, such as the number of training days or a suggested program, so the end of the flow has something concrete to show.
The coach's voice sets the tone
Aram's tone in the app is hard and direct, "man to man", without humiliating anyone. In onboarding that line matters most. A first-time user who feels mocked closes the app, while one who feels challenged and respected keeps tapping.
Write the tone as a rule sheet before the script, then check every clip and every button label against it:
| Rule | Weak button | Button that follows the rule |
|---|---|---|
| Speak to the user as "you" | "Option B" | "I train three times a week" |
| Never comment on the body | "I'm out of shape" | "I'm starting from zero" |
| Praise the decision, promise no result | "Lose 10 kg" | "I'm ready to start this week" |
Other businesses need other rules. The AI assistant on Lucent Legal works under a written rule to say it is an assistant, not a lawyer, and to never pose as a person. A rule that short is easy to check line by line. A clinic might use "warm, no medical promises". A law firm might use "calm, precise, no jargon in buttons".
Ask for the name, then use it
Imperia's onboarding includes a step where the user enters their name. Typing a name is a small commitment, and after it the app can address the person directly.
How to set up the step:
- Place it after the first exchange, once the user has seen the app respond. A name field on screen one reads like a form.
- Use one field for the first name only, with no surname and no validation that rejects Cyrillic or Arabic script. A UAE audience types in all three.
- Use the name within the next one or two screens, in the coach's line or the button text, and again on the home screen.
- Make the field skippable and have a fallback greeting ready.
If you ask for the name and never use it, you have spent one step of the user's patience and got nothing back.
Build the engine once, reuse it
The film player in Imperia is one engine, OnbFilm. It drives both the onboarding and THE RUN, the running program, where a voice tells the user by heart-rate zone when to run and when to walk. Both features need the same parts:
- clip playback with preload, so the next clip starts without a pause
- prompts with buttons over video
- branching rules: which answer leads to which clip
- voice cues and timing
If a developer builds the onboarding as a one-off, you pay again for every later feature with interactive video, such as a tutorial, a guided session or a product walkthrough. At scoping, ask one question: "Will the onboarding player run other screens, and what does adding a new film cost after launch?" Scope and prices for app work are on the apps service page.
The paywall comes after value
Imperia has two plans, Basic and Pro, with 3 days free, and the mentor chat sits in Pro. The onboarding has to make that choice clear: what each plan gives and what happens when the free days end.
Whatever your pricing, the order stays the same. First the user sees something shaped by their answers. Then you show the plans in plain words, with the trial terms on the same screen. If the paywall appears before anything personal, the user has nothing to judge the price against.
Apple reviews how subscriptions are presented, so the wording on this screen needs care. More in what Apple checks in review.
Measure every step separately
One completion number tells you nothing about what to fix. Log an event when each screen opens and when each answer is chosen, with the branch name attached. That shows the exact clip where people leave.
| What to log | Why it matters |
|---|---|
| Screen shown, with step number | Shows where the drop-off sits |
| Answer chosen, with branch ID | Shows which branches lose people |
| Clip watched to the end or skipped | Shows a clip that runs too long |
| Name entered | First commitment point |
| Paywall shown, trial started | Connects onboarding to revenue |
| First key action after onboarding | Shows whether the flow led anywhere |
If users arrive from ads or your site, tag the links with UTM parameters to compare how each source moves through the flow.
Language and audience
Imperia speaks to a Russian-speaking audience worldwide. The interface is in Russian, and the App Store listing uses one text across all locales. A UAE business usually has a mixed audience, and a film onboarding makes each extra language more expensive:
| Option | What you produce | Trade-off |
|---|---|---|
| One language, subtitles in others | One set of clips plus subtitle files | Cheapest to add; the voice stays in one language |
| Voiceover per language | One shoot plus a recorded voice track per language | The coach's own voice is lost |
| Separate takes per language | A full shoot per language | Highest cost; only if the speaker is fluent |
In every option, write the button text natively in each language, because buttons are cheap to change. For Arabic, check that the button layout mirrors right to left. Decide all of this before the shoot. Re-filming after launch costs more than planning it in. For a venue app, the full path from idea to submission is in a venue app from idea to App Store.
What not to do
- Don't open with a sign-up form. Let the user see the product respond before you ask for an email or a phone number.
- Don't ask questions you won't use. Every unused question lowers the chance the user reaches the end.
- Don't film the coach reading a feature list. The App Store page already lists features; the film should answer the user.
- Don't make clips that can't be skipped or replayed. People open apps in lifts and queues with the sound off, so add subtitles.
- Don't branch without merging. A tree with eight endings means eight times the edit and eight times the testing.
- Don't hide the trial terms. If the price and the end date of the free period are missing from the paywall screen, users feel tricked and review may push back.
- Don't launch without per-step events.
Checklist
- List what the app must know after onboarding and the one action that follows it.
- For each question, write what changes on the next screen.
- Write tone rules with button examples before the script.
- Limit each question to two to four button answers and write a reply clip for each.
- Plan where branches merge.
- Put the name step after the first exchange, accept any script, use the name within two screens.
- Ask the developer whether the onboarding player can run other screens.
- Show plans and trial terms only after the user has seen something personal.
- Add subtitles and a skip option to every clip.
- Choose subtitles, voiceover or separate takes before filming.
- Log screen, answer, clip, paywall and first action as separate events.
- After launch, review drop-off per branch and re-cut the clip where most people leave. The app itself is on the IMPERIA THE1 project page.
Questions and answers
How long should an app onboarding be?
Long enough to collect what the first session needs, and no longer. Each question should change something the user sees next. If a question only feeds a database, move it to the profile screen.
Is a video onboarding worth the production cost?
It makes sense when a person is the product: a coach, a doctor, a lawyer, a host. If your app is a utility, three clear screens will do the job at a fraction of the cost.
Where should the paywall go?
After the user has seen something built around their answers, with a plain description of what the trial includes. In IMPERIA THE1 there are two plans, Basic and Pro, with 3 days free, and mentor chat comes with Pro.
How much does an app with this kind of onboarding cost?
My app work starts from AED 35,000 (Special offer). A film-style onboarding adds a shoot and an edit on top of the code, so fix the scope before the first take.