The store did not have a growth problem. It had a plumbing problem. Subscriptions were signing up fine and leaking out the back through failed payments nobody retried.

We started with dunning. A card declines for a hundred reasons, most of them temporary, and a single retry loses revenue that three well-timed retries recover. We wired a schedule around the network's own retry windows and watched involuntary churn drop by a third.

Then the edge cases: paused plans, prorated upgrades, gift subscriptions that should not renew. Each was a branch the old flow never handled, and each was quietly cancelling someone.

Retention is rarely a marketing fix. Here it was a billing fix, shipped as a plugin.