Jonathan Parra - I made 4,000 app paywalls and learned this

Control: plans ordered weekly/annual/monthly with long names ('AI Assistant Weekly', 'AI Assistant Annual'), a small 'Limited time 70% off' badge and a CTA 'Put AI into action'. Variant: plans renamed 'Annual plan', 'Monthly plan', 'Weekly plan' with annual on top and highlighted, price compared against the cheapest plan, badge cut to '70% off' and centered ('limited time' in small text doesn't work; put urgency in the heading only on a true one-time-offer paywall), a check mark 'No commitment, cancel anytime' line, and a plain 'Continue' button.

Removed the monthly plan so the picker shows annual (still preselected, since it has the highest LTV and ARPU) and weekly with a 3-day free trial; CTA changed to 'Start my 3-day free trial' instead of 'Put AI into action'. Fewer options cut decision fatigue and the trial copy makes users feel safe continuing. Guest would also test an annual + monthly variant, and puts any third plan behind a 'View all plans' button.

Add a 'No commitment, cancel anytime' subtitle or check-mark line near the CTA on nearly every paywall the guest designs; described as a simple hack with a granular bump in conversion rate.

Avoid showing three plans on the main footer (decision fatigue); show two (e.g. annual + weekly) and hide others behind a 'View all plans' button, or sometimes hide the whole product group. When several plans are shown, order them strictly by length (ascending or descending), never weekly/annual/monthly. Default-select yearly because it usually has the highest LTV.

Main paywall defaults to the ~$39-40/year plan with 'Start' and small price text. On close, a drawer slides up: 'Not ready to commit for a year? We have plans for everyone.' Yearly stays selected but is shown as a weekly price ($0.76/week) next to a monthly or weekly plan. If users tap 'Not now, thanks', show a one-time-offer paywall with the yearly plan at ~$30/year (33% off) and a chevron arrow on the button. If many users pick monthly, read it as a sign onboarding didn't show the product well enough.

For apps that transform the user over a week, month or other timeline, use a Clear 30-inspired paywall with steps showing where the user will be after the plan, positioning the user as the hero of their story; add 'No commitment, cancel anytime' and default to yearly. Best for apps with a clear USP and a closed feedback loop (before vs after).

Use 25-33% off on regular one-time-offer paywalls. Save large discounts for Black Friday / Cyber Monday; running too many or too-large sales trains freemium users to wait for deals and cheapens the product, so avoid big discounts especially for larger apps.

Control: 'Upgrade from free to premium' with a free-vs-premium comparison table and 3/6/12-month plans with dense price copy ('$49.99, not $99.99, monthly price billed quarterly'). Variant (copied from Cal AI): headline 'We want you to try [app name] for 50% off' (or '...for free' when a trial plan is selected), a large image or video of the app on a phone, 'No commitment, cancel anytime', and a bigger 'Continue' button leading to a second screen with more details.

Default CTA height 65pt (sometimes 56pt) with large button text; a right chevron arrow on the button also bumps conversions (tested as its own variant, not bundled with other changes). Tip came from Superwall's founder Jake in an internal channel and was rolled out across paywalls.

First paywall for a new app: one screen with the USP heading and subtitle, a bullet list of what you get, social accolades, a 'No commitment' subtitle and the continue button. Use it as the baseline, then build variants by dropping in reusable components: an image or video of the app in action (Fitness AI style), extra pages with social proof/reviews for larger apps, a 'Try 7 days free' CTA with chevron, a table instead of bullets, etc.

Prioritize design tests and price-packaging tests (which plans, naming, order). Delay price-point tests: in Superwall they require keeping users in fixed cohorts/audiences via user seeds for every experiment, which is painful to manage, so run them only after many other tests.

Mojo-inspired paywall: a Rotato-rendered video of the app in action on a phone (guest screen-records himself using the app and drops it into a Rotato template), social proof, simple bullets, monthly vs yearly plans, and a targeted CTA 'Start my glow up' for a largely female audience. Components (video, slider, text) are then swapped in further variants.

For an app built on stock SwiftUI styling, designed an Apple-esque paywall (app icon at top, SF Pro / SF Pro Rounded text, SwiftUI list-item styles) and tested it against more customized variants. Guest repeated it with another app. A Superwall template exists for this style.

Two-tier app (Plus and higher Pro, each weekly/monthly/quarterly), Hinge-inspired paywall. The Plus tier shows on a white background; selecting Pro fades to a dark premium-looking background, contrasting with the onboarding so users sense something is different. Users toggle between tiers on the paywall and land back on the premium look.

Paywall lays out the trial as a timeline: today you unlock the app, day 5 you get a reminder notification, day 7 billing starts. Widely copied from Blinkist.

A phone chat product with a carefully tuned, personalized prompt for one niche, billed at $20-22/month (ChatGPT/Claude level) instead of the ~$30-40/year typical of AI wrappers.