What actually works — and why.

Principle 01

Show the outcome. Never teach.

The strongest, best-evidenced finding in the whole field. NN/g's study found task success was 94% for people who skipped the tutorial vs 91% for those who read it — no gain — and, counterintuitively, tutorial-readers rated the same tasks as harder. Their blunt conclusion: "avoid onboarding whenever possible; keep any instruction brief and optional."

Principle 02

Get to the aha in the first session.

RevenueCat's data across 75k+ apps: ~82% of trials start on day 0, and 55% of 3-day-trial cancellations happen on day 0. If the user doesn't feel value in the first session, they're gone. Onboarding is the retention mechanism — not a preamble to it.

Principle 03

Animate the magic — over a real screenshot.

The demo beat is the hook. The validated pattern (NN/g / Material "Top User Benefits"): autoplay with a manual-advance override — the animation plays itself (photo button taps, photo taken, item chips pop in), pagination dots show more exists, and the moment a user touches it, autoplay pauses so they can flip. One action per screen. Play it over the real app UI, not an abstract illustration — which is exactly what the live mockups in this lab do.

Principle 04

Cap standalone benefits at three. Branch the rest.

Material caps the auto-carousel at ≤3 primary benefits — more dilutes the message. NoonFox has more than three great use-cases (passports/keys, expirations, warranties, storage). So either pick the top 2–3 by emotional pull for standalone screens and fold the rest into one "…and everything else" grid, or branch to the matching use-case from the segmentation question (the zero-party-data pattern that personalizes and deepens investment).

Principle 05

Put value before the signup ask.

Duolingo's "product before signup" — letting users complete a lesson before any account ask — lifted DAU ~20%. NoonFox analog: let people see (or even take) a snap before the sign-in/paywall ask. Motivation should peak first; the form comes when the user already wants what's behind it.

Principle 06

Time permissions by context, and say why.

Camera is core → ask early, with a one-line reason before the OS dialog (any reason lifts grants ~12%; a good one up to ~81% in an A/B test). Notifications & location → defer to the moment they're needed. Never blast all permissions at once. iOS gives one shot — a premature "Don't Allow" is permanent, so prime it.