Case Study: Architecting a Stripe v2-to-v3 Payments Bridge at Jetty

The Business Problem

At Jetty, we had an underperforming product, Jetty Credit, with a 0.42% take rate. Leadership needed it to hit a 5% adoption rate to justify its continued development. The only viable path was to cross-sell it during our main Jetty Deposit checkout flow. This was a high-stakes problem: Jetty Deposit was our "golden goose," driving 70% of the company's revenue, and we could not risk its stability.

The Core Technical Challenge

This wasn't a simple UI fix; it was an architectural nightmare.

• Jetty Deposit (Core Revenue): Ran on deprecated Stripe v2 (Sources).

• Jetty Credit (New Product): Ran on modern Stripe v3 (PaymentMethods).

These two systems are fundamentally incompatible. You can't use a v2 Source and a v3 PaymentMethod in a single unified flow. The "obvious" fix, migrating our entire core revenue platform to Stripe v3, was a 6+ month, high-risk project that no one was willing to take on. We needed another way.

My Approach & Solution

I led the technical design for a "payments bridge" to solve this. My entire philosophy was to de-risk the project and avoid the migration.

Instead of changing the legacy system, I architected a dual-flow architecture that would, from a single user input at checkout:

• Create a legacy Stripe v2 Source to process the immediate Jetty Deposit payment. (This kept the core revenue 100% isolated and safe).

• Simultaneously create a modern Stripe v3 SetupIntent to authorize future billing for Jetty Credit. (This deferred all the billing complexity and avoided refund/charge issues).

The whole integration was heavily feature-flagged and built with deep Sentry and Datadog monitoring from day one, so we could roll it back instantly if we saw any issues.

The Outcome (The Factual Result)

The architecture worked. We ran a painted door test that proved a 74% checkout completion rate, validating the flow. This gave leadership the proof they needed that we could hit the 5% adoption target and unlock a new revenue stream (projected at $2.4M ARR).

Technically, the achievement was that we successfully bridged two incompatible Stripe systems in a live, high-traffic, SOC2-compliant environment, with zero degradation to the core checkout flow. It was the right, pragmatic solution that delivered the business goal without the 6-month, high-risk migration.