A verdict on your three proposals (a short-window impulse pass, the "broken" Stripe checkout, and price-stating button copy), the overlooked variants, and the order to ship them in.
The pass is the right direction (it fits one-shot Google intent far better than a recurring subscription). All four converged on the same shape. But two of your beliefs need correcting, and a cheaper first door exists.
Your instinct was that people reach the payment page and back out en masse. The funnel says the loss is almost entirely one step earlier.
1,140 unique people hit the paywall. 1,115 never even clicked toward a checkout. Only ~25 reached Stripe at all — and that tiny number is polluted by your own test purchases and unmeasured (checkout_completed has never fired). So "the checkout is where they bail" is almost certainly a misattribution. The wall itself, and the ask behind it, is the leak.
Before touching a price or a button: (1) fire checkout_completed server-side from the Stripe webhook, (2) exclude your own account from every funnel count, (3) tag each purchase as pass vs subscription. Ship two changes without this and you will have moved two variables and still be blind.
The agreed shape, and the two traps to avoid.
Hoy sin límite · 1€, labelled "sin suscripción", kept visually secondary to the subscription so it does not cheapen it. ("Pase" is new vocabulary — have the copy editor bless it against the naming/vocabulary law.)A €1/week auto-renewing pass is €4.33/month, which is more expensive than your €3/month plan. That trains your best prospects to overpay or churn. The pass must be one-shot and expire on its own.
Stripe takes about €0.25 + 1.5% per charge. The fixed fee is what kills sub-€1 prices.
Build cost, honestly: a real time-boxed pass is a few days of backend work (a new one-time payment-mode product, a passes table with expires_at, an expiry-aware access check, and webhook fulfilment). The existing sponsor checkout already does one-time payments, so the pattern exists to copy.
Right now the reader falls from a rich, reassuring card onto a bare Stripe page. Closing that gap is a small change in one adapter plus two dashboard edits. Do it regardless of the pass.
locale = es (today it defaults to the browser language, so it can render in English).custom_text that repeats the three benefits (Todas las secciones · Sin límite diario · Todo el archivo) and "Pago seguro" right by the pay button.Ranked by leverage-per-effort. The top two are cheaper than the pass and aim at the real leak.
A Stripe coupon that makes the first month €1, then €3/mo. Near-zero build (no new entitlement model), gives you the same impulse anchor and recurring revenue. John's pick to test before building the pass.
Cheapest recurring-revenue doorShe arrived from Google for ONE notice. Killing it mid-paragraph betrays that promise. Let her finish this one, then wall the depth (the second article, the archive, other sections). Convert on satisfaction, not hostage-taking.
UX change, reversibleBuild the habit before asking for money; watch whether return-visits and paywall click-through rise. Cheap, reversible, measurable.
Config changeSurface an Apple Pay / Bizum express button on your own card, so wallet users never see the Stripe page at all. Removes the context-drop entirely for the highest-intent buyers.
Small buildFor the 1,115 who will never pay today: an "te lo guardo" / free alert or digest signup as the first conversion. Turn one anonymous visit into a relationship, then monetize the ongoing job (alerts for "your" oposición / subvención).
Reclaims the 97.8%checkout_completed from the webhook, exclude your own subs, tag pass vs subscription. ~1 day. Nothing below is readable without it.Pass is fine but aimed one step past the leak. Ship a "primer mes a 1€" coupon first. Pass: €1, 24h, one-time, site-wide.
The bare Stripe page is a real emotional drop — fix continuity. But don't hard-gate the article she came for: soft-wall the depth. Button: action-first, not price-restating.
Cost asymmetry dictates order: checkout enrichment is hours, the pass is days + a migration. Instrument first. One-time payment, a passes table, never auto-renew.
"Checkout is broken" is misattribution — 97.8% is lost before Stripe. €1 is the fee floor. Win = ≥1% of paywall viewers buying within 4 weeks.