Personalization tactics

Modeled on Clear30's onboarding and Alex Hormozi's value equation (raise dream outcome and perceived likelihood, cut time delay and effort). Flow: 'Let us build your program', enter name, pick goals from many options (dream outcome), how much time you waste daily (pain), which apps are most addictive (later screens show that app), how it makes you feel (show the opposite feeling), 'What have you already tried?' followed by a screen explaining why each method (e.g. screen time limiters) fails, with tappable links to cited research, positioning the app as the best way to quit. Social proof screens sit between questions, and extra trust-building before the screen time and notification permission screens where many drop. A shorter version is being tested because the paywall rate is lower than they want.

Three tiers (Silver, Gold, Platinum); most users pick Gold and the paywall can't explain everything, so it normally leads with Gold. A machine-learning model using ~900 data points per user (profile, behavior, previous experiment results) predicts propensity to buy Platinum (wider roadside assistance range, identity protection, more services). High-propensity users are shown Platinum instead of Gold. Part of an ML targeting/orchestration layer built in the last year that also targets promotions, feature education, push notifications and in-app messages.

After iterating one or two onboarding screens at a time (small wins, hard to make big progress), they made one big jump: from ~5 to ~13 screens ending in the paywall. Added screens include login, notification permissions and mostly personalization (what the user is studying). Login was moved to the last screen, after the paywall.

Test roughly doubled onboarding length (to ~15 screens) with more personalization questions and social proof, then paywall; aim is a trial start in the first session.

Onboarding asks the user's goal (weight loss, getting started, having fun, reducing stress). The app then builds a goal-based program (a curated package of classes) and drops them into it right away instead of a content catalog, to cut decision paralysis. Framed against free YouTube: 'how is this better than YouTube?' Kept releasing more programs and features that reduce choice.

About 90% of traffic goes to a long personalized quiz on the website. The quiz educates users on the product, overcomes objections, raises perceived value and shows how personalized the plan will be; answers are used to build the user's plan in the app. Tested traffic to the App Store vs the quiz: the quiz converted much higher. An almost identical quiz inside the iOS app converted worse than web, while the same quiz on Android converted about as well as web. They may retest App Store traffic later.

After tearing down Cal AI, Quitter and Ahead onboardings: intro screen shows the dream outcome (30 days clear-headed), not features. Ask their name, then use it in the next question. Ask short- and long-term goals, then reflect them back with social proof ('81% of users saw improved mental clarity'). Ask pain questions (days a week, money spent), then compare them to others with cited sources. Then show where they'll be in 30 days (money saved), that the program does the work, that 25,000+ people did it, and that users improve within the first 3 days. Frames every slide on the value equation: dream outcome x likelihood / (time delay x effort).

Problem: many tiers (Plus/Gold/Platinum) x plans (weekly/monthly) plus a la carte items caused decision overload; some users bought nothing, others bought Platinum but only used Gold features. The ML team built models predicting willingness to pay; at paywall time the paywall asks the model which paywall/product to recommend for that user, instead of showing everyone the same static paywall. Rolled out small first (a couple of features/paywalls, not the whole scope) to limit cost and risk. A/B test: control = historical static paywall, treatment = model-chosen paywall; measured conversion and total revenue. Counter-metrics: repurchase, renewal and cancellation of users nudged to higher tiers; several model variants compared against each other and against no-model paywall.

Replaced a rush-to-paywall flow with a longer onboarding that asks about interests and fitness level, shows value interstitials, and recommends a program before the user sees the app. Answers seed the recommendation algorithm, whether onboarding starts on web or app.

While analysis runs, a 'warming up' screen shows how much data is being processed. Then: minutes of stress from a recent clean day (e.g. 192 minutes), a timeline of their day with stressed/active/recovering periods and an animated zoom on a blue recovery period ('your body was doing a good job' - empowerment after the scare), then an animated 30-day calendar of built-up stress, and the share they relieved on their own (~20-30%). All numbers come from the user's real Apple Watch data.

Ad story continues into the funnel: users arriving from an ad get a long question flow (work situation, tools, where they chat, pain points, goals) with answers reflected back and a summary at the end, then 'Get Sam for $1' promo week, then $10/week, straight to Stripe checkout. Answers are actually used to personalize the agent. Direct website visitors skip most questions because they weren't warmed up by an ad. Funnel length and screens are A/B tested.

Ask age in onboarding. Under 25: short flow (key points, disclaimers to accept, straight into the app). 25+: longer flow on nutrition/sleep/water that builds a personal nutrition plan (height, weight -> recommended calories/protein) to convince them tracking matters.