Sometimes an ugly paywall outperforms a beautiful one, and the operators who tested thousands of them say the lesson is about sequence and evidence, not polish. A paywall is the visible end of a decision the user has often already started making. The three themes below follow that arc: the paywall as a flow, the levers that reduce risk or add friction, and the metrics that decide which result counts. Treat the paywall as a flow, not a screen Sequence and the decision before the decision Jonathan, who says he has studied nearly 3,000 paywalls, argues that multi-page paywalls almost always beat single-page ones because each page lets the user absorb one claim at a time. He pairs that structural point with timing: some users decide to pay before they ever see the paywall, so onboarding, feature gates, settings screens, and win-back offers all carry the decision. The Moonly founder makes the same argument from the product side. Their paywall opens by selling the outcome, eight years of life back, and once that is shown the screen stops feeling like an interruption. Showing users how many years they could save lifted trial sign-ups from 7% to 17%. That is one app's before-and-after figure, and the material does not report sample sizes or whether the lift held. Multi-page paywalls almost always do better than single-page paywalls. Levers that reduce risk, and the friction that filters Reducing perceived risk Blinkist's users said they had been tricked into being charged, so the team replaced the trial description with a step-by-step timeline. Trial sign-ups rose, complaints fell, and push notification opt-ins also went up. Tipstop kept the offer identical and changed only the presentation, adding a discount badge and clearer framing. Direct conversions almost tripled. Slopes borrowed the logic of Apple's redeem flow for a single-action 'pay ramp' that raised trial starts by 25%. Each of these cases attacks the same hesitation the source names directly: fear of forgetting to cancel, and fear of choosing wrong. When friction helps Friction is the counterintuitive case. Outsider required a credit card to start its trial. Sign-ups dropped by more than half, but conversion rose fivefold, more than doubling paying customers. Moonly offered its free trial only on the annual plan, a similar filter. The explanation offered is that the extra step removes people who were never serious. The material does not report revenue per visitor, so the friction claim rests on conversion and paying-customer counts alone. What to measure, and where the sources disagree Metrics are where the sources disagree most. The paywall designer warns that ARPU can mislead: a three to four dollar ARPU may come from weekly products with 90 to 95% monthly churn. He tracks paywall percentage, trial-to-paid conversion, LTV, D1, D7, D30, and churn, and says the right weighting depends on the business. Moonly's founder argues from the opposite direction on pricing. A high-priced lifetime plan can anchor users toward another option, and switching units from monthly to weekly brought money in faster. Both are reporting real results from different businesses, and neither transfers without its context. The spin-the-wheel dispute is sharper still. The designer rejects predetermined wheels that land on discounts, while conceding that they maximize revenue on weekly prices. Quick revenue and a durable subscriber base pull against each other here, and the sources do not resolve that tension. Test radically different paywall designs before micro-tuning copy, because the source reports that design tests move results the most. Show the outcome before the price, and quantify it where you can, as the 7% to 17% trial lift in one app suggests. Pair every trial with a visible timeline or cancel-anytime cue, since Blinkist's complaint fell after it added one. Treat friction such as a card requirement as a filter to measure, not a default to avoid. Track paying customers, not only sign-ups. Read ARPU alongside churn, trial-to-paid conversion, and LTV before deciding a paywall has won. The capability this chapter promises is a way to design a paywall as a sequence of decisions you can test. Start with the flow and its timing, choose the risk-reducing or filtering lever that fits your business, and judge each variant by retention and LTV rather than by how polished it looks. Explore the paywall with the questions the sources keep returning to: which business are you building, and what evidence would make you abandon this design?