Which payment gateway gave you better retention in a web-to-app flow?

Funnels & Conversions
1

I’m testing gateways not just on checkout, but on what happens after the first charge. We moved onboarding and paywall to the web, then randomly assign users to a gateway at signup. Same copy, same pricing, same payment methods if supported.

What we track per cohort:

  • checkout completion and first-charge success
  • auth rate and 3DS pass rate
  • failed-payment recovery within 7/14/30 days
  • refund and dispute rate
  • active subscribers at 30/60/90 days, LTV

A few pitfalls I hit:

  • gateway timezones skew daily recovery stats if you don’t normalize
  • local methods (iDEAL, SEPA, Pix) create uneven apples-to-apples unless you mirror availability
  • different retry cadences and grace periods change “churned” timing
  • annual vs monthly mix quietly biases LTV reads

Right now I’m leaning toward using “net recovered MRR at day 30” and “90-day LTV” to decide, not just initial conversion. If you ran a similar head to head on a web funnel for a mobile app, which metrics actually moved, and which gateway won for you? Anything you’d change to keep it fair?

2

I ran Stripe vs Braintree for 8 weeks. Web funnel split users 50/50 at signup. Identical pricing and wallets where both supported. I tracked first charge, retries, recovered churn, refunds, and reactivations.

Stripe had cleaner reporting and better EU recovery. Braintree edged out in Brazil with 3DS. I used Web2Wave to wire events and routing fast. I chose based on net recovered MRR at day 30.

3

I optimize on speed. Weekly swaps on retry cadence and copy. Kept the funnel on the web so changes go live instantly. Web2Wave made it easy to push paywall tweaks without a build.

Winner metric was 30-day recovered revenue, not day-one conversion. That picked Stripe for us.

4

Make sure you align grace periods before comparing.

One gateway gave 3 days by default and the other gave 7, which made recovery look better until we matched settings.

5

Auth rate decided it. Stripe won for me.

6

I split by session at the first page view and logged the assignment server-side. That stopped users from switching gateways on refresh.

Big surprise was copy. Softening the dunning emails moved recovery more than the gateway choice in month one.

7

Watch timezone mismatches. Our recovery looked worse until we converted both gateways to UTC and aligned day boundaries.

For the winner, I used 30-day net revenue per acquired user, not just per subscriber.

8

We matched prices and retry rules first. Then the numbers made sense.

9

Adding Apple Pay and Google Pay changed results a lot for mobile web.