Paywalls picked by demographic: +83% MRR in 30 days
Commentary byLiftArchiveOur analysis, not theirsFlirt AI grew MRR 83% in 30 days after showing paywalls chosen by demographic data from onboarding, with trial-to-paid up 20%.
The AI dating reply app ran three paywall variants at once and shipped product and pricing changes at the same time.
The result was published as a Superwall case study.
- MRR: +83% within 30 days.
- Trial-to-paid conversion: +20%.
- These figures cover the paywall test plus product and pricing changes together, so no single change can claim them.
Control
The app's earlier paywall, shown to all users.
Variant
Three paywall designs with different copy, imagery and CTAs, with the one shown picked by demographic data collected in onboarding. Product and pricing changes shipped alongside.
Why it likely won
Our reading of the result, not the practitioner's claim.
- Copy and imagery aimed at the core audience may have made the value clearer to the people most likely to pay.
- Product and pricing changes likely added to the revenue gain on their own.
- Higher trial-to-paid suggests the new paywalls drew more committed trial users, not just more of them.
Caveats: This was bundled: paywall targeting, design, product and pricing all changed, and the MRR figure covers 30 days of growth rather than a clean split. No sample size or significance was shared. Superwall sells the paywall and testing tool used here, and its published case studies can lean positive.
Who ran it
Flirt AI ran this test. Source: Superwall case study. The numbers are theirs; the summary and commentary on this page are ours.
How to run this test on your app
- Hypothesis
- If we use an onboarding answer to show each group a paywall with copy and imagery made for them, then trial-to-paid will rise because the offer will match what each group wants.
- Primary metric
- Trial-to-paid conversion per segment
- Guardrails
- Trial start rate
- Proceeds per user
- Refund rate
- Who should test it
- Apps that already ask a demographic or goal question in onboarding and have a few clearly different user groups. It is less useful if your audience is uniform or traffic is too low to split three ways.
Questions
Can we credit the 83% MRR growth to targeted paywalls?
No. Product and pricing changes shipped at the same time, so the figure reflects the whole package.
What data should drive paywall targeting?
Flirt AI used demographic answers from onboarding. Any answer that splits users into groups with different needs can work, as long as each group gets enough traffic.
How many targeted variants should we run?
Flirt AI ran three. More variants split your traffic further, so match the number to how many users reach the paywall.
Similar tests
Kovalee