The Subscription Growth Formula: Churn Math, Retention Wins, and Smart Product Bets — Dan Layfield, Subscription Index

Downloaded all checkout page payment errors, stack-ranked them by frequency, rewrote the copy so each failure tells the user what is wrong and what to do (instead of 'Invalid payment method. Please contact your bank.'), with fallbacks: if the card fails, offer PayPal; if PayPal fails, send them back to a card. Merged back into the codebase. Took about 2 days. Rationale: 100% of new revenue passes through checkout, so a 1% lift lifts all revenue 1%.

Cancellation survey routes each reason to a save offer: 'too busy' -> pause the subscription (worked best at Codecademy since people learn to code in bursts); 'too expensive' -> temporary discount (same plan at 50% off for 3 months, then auto-return to full price) or a downgrade to a lower, hidden tier; 'negative experience' -> connect with customer support. On Stripe this needs custom logic in the payment system. Watch the share of users who take the discounts, since save tactics become widely known and get gamed.

Grouped countries into about three willingness-to-pay tiers (Nordics, Western Europe, US and Canada highest; a second tier paying less; a third paying least) and localized price level as well as currency and page content, initially pegging FX rates. Cost: 10 price packages to manage in ads, email, support docs and every future price change. Guest advises squeezing easier tactics out first.

Set up dunning/email notifications when recurring payments fail and spend about 2 days copywriting them, since they reach 100% of users with failed payments. On the App Store, detect billing issues where possible and send custom emails asking users to update their payment method. The lower churn gets, the larger payment-failure churn's share (at 2% churn, maybe half).

For web checkout, measure the share of 'pay' clicks where money is actually collected (almost no one tracks this at first). Dig into failure reasons: US companies charging worldwide through only Stripe or US gateways lose payments in regions like India and China; as you grow, add secondary payment providers or a subscription manager. Less relevant for App Store apps.

Right after purchase, ask a single open-ended question such as 'Was there anything about the purchase experience that didn't work for you? Was it confusing?' or 'Were there any questions about this product you couldn't find an answer to before purchase?'. Combine with fresh-eyes walkthroughs (friends, UserTesting.com) and work the funnel bottom-up from the highest-intent users.

When a user taps the paywall CTA and then cancels the App Store payment sheet, a modal pops up offering a reverse free trial: 7 days of premium completely free that does not auto-renew.

After a user taps the paywall CTA and dismisses the App Store payment sheet, immediately show a discount offer (e.g. 50% off) to win them back. Host notes Apple has reportedly hassled some developers over this as a form of price discrimination.

Host's weather app spent 5 months on one big innovative feature, hoping it would get press coverage and work well in marketing (TikTok videos, Meta ads). No dollar ROI estimate and no validation (e.g. mockup tests on TikTok) was done beforehand.

Large subscription companies (The New York Times, CLEAR, Zoom) use cancel-survey answers to trigger save offers: 'too expensive' -> temporary two-month discount, bad experience -> support, too busy -> pause. Zoom offered the guest, paying ~$15/month, a hidden ~$7/month plan with a limited number of calls.

In a long-term use case with easy switching (food delivery, rides), sell a membership (~$9.99/month for Uber Eats in the US) as a retention tool. Since the core business makes the money, the subscription can be priced flexibly; its job is to stop switching. Used by Uber Eats, Uber, DoorDash and Lyft.