What this prompt does
This prompt treats onboarding as engineering, building a flow designed to maximize activation. It covers six steps: mapping the flow as a state machine, building animated welcome screens, implementing pre-permission priming, creating a personalization survey, adding a delayed registration wall, and setting up A/B testing. Mapping the full state machine first is what prevents the common mistake of begging for an account on screen one, before the user has any reason to commit.
The variables shape the flow to your app. [step_count] and [survey_questions] size the journey, [ui_framework] and [carousel_count] plus [animation_type] drive the welcome screens, and [permission_type] with [remind_delay] control the priming. [goal_options] and [storage] personalize the experience, [registration_gate] decides when the account wall appears, and [ab_test_tool], [variant_count], and [test_duration] configure experimentation. Each variable maps to a decision that directly affects activation, so they are worth tuning rather than leaving default.
When to use it
- You are building a new app's onboarding and want it tuned for activation, not decoration.
- You want the full flow modeled as a state machine before any screens are built.
- You need pre-permission priming so
[permission_type]grant rates improve. - You want a personalization survey that actually customizes the first session.
- You want to delay the account wall until the user has built something worth saving.
- You plan to A/B test step count, copy, and permission timing rather than argue about them.
Example output
Expect a state diagram covering every path including skip, back, and conditional branches based on survey answers. The welcome step produces animated slide specs with headlines, subtext, and CTAs; the permission step produces priming screens with grant/deny tracking and a remind-later option; and the survey step produces a [survey_questions]-question flow using visual inputs. Later steps add a delayed registration step gated at [registration_gate] and an A/B test setup with [variant_count] variants in [ab_test_tool] tracking completion rate and D1 retention. It mixes flow diagrams, UI specs, and tracking definitions, with the conditional branches drawn out so no path is left unhandled.
Pro tips
- Get the state machine right first; set
[step_count]and map skip and back paths before building, since rework here is expensive. - Use pre-permission priming for
[permission_type]and offer a remind-later that resurfaces after[remind_delay]; never trigger the system dialog cold. - Keep
[survey_questions]short and visual; long surveys with standard form fields kill completion rates. - Place the
[registration_gate]where the user has already invested effort, and surface that progress as motivation to sign up. - Store personalization in
[storage]so the home screen genuinely reflects the chosen[goal_options]on first launch. - Run each A/B variant for the full
[test_duration]; calling a winner early on activation metrics is misleading and easy to do.
