Rebuilding Recurring Buy around what users would actually delegate

• Users didn't trust the system yet

  A recurring plan isn't a one-time purchase. A purchase ends. A plan keeps making decisions after the user leaves. At launch we were asking people to schedule future purchases before the system had earned that. The commitment felt larger than the interface let on.

• Users couldn't invest the way they normally would

  Many users wanted to fund recurring purchases using ACH and stablecoins. Neither existed in V1. This wasn't a UX problem — it was an infrastructure problem. Recognizing that early kept us from trying to patch missing payment rails with better screens.

• Users couldn't tell whether a plan was working

  The product went quiet after setup. Users could see a plan existed, but not whether it was healthy, whether the next purchase would execute, or how to adjust without starting over. When money is involved, silence reads as risk.

The principle I set for V2: remove repetitive effort, preserve judgment. Not maximum automation — the right amount of it. People wanted help executing their investing strategy. They didn't want the product replacing it. In parallel, we closed the infrastructure gap behind failure two — ACH and stablecoin funding shipped alongside V2, so better screens weren't papering over missing rails.

1

Recurring buying became part of buying

We stopped treating recurrence as a separate product. Instead, it became a choice inside the buy flow people already used. Buy once, or buy on a schedule. That's it. The decision got smaller. It also dissolved the funnel we’d been measuring — once recurrence became a choice inside an existing flow, a standalone “completion rate” stopped existing; volume and active plans became the honest scoreboard.

2

Plans became easier to understand

The redesigned plan view answered the first question users actually had: “Is this working?” Average entry price. Current price. Total invested. Returns. Next execution. Visible at a glance.

3

Failures became impossible to miss

V1 let plans fail quietly. V2 didn't. Low-balance warnings surfaced sooner. Repeated failures paused plans automatically. Removing a payment method previewed future impact before you confirmed anything. Recovery stopped feeling like detective work.

4

Automation became accountable

We surfaced recurring purchases in the same places users already tracked their money: Portfolio, Activity, Transaction history. Hidden automation feels suspicious. Automation that shows up in public feels like it's working for you.

That changed the roadmap — I made the case for it. Instead of more automation, we gave users a better way to express intent. They defined a price range; the system executed purchases only inside it, handling all the monitoring in the background. The user kept the rule.

V3 shipped latest in the window, so its adoption numbers are still maturing — and the result I trust most didn't show up during launch week anyway. It showed up during the first real market correction: one-time buying dropped, recurring buying held. It’s also the cleanest control I have on the growth numbers above: if 5× volume had been bull-market beta, both lines would have moved together. That was the original promise finally showing up in production — the system staying consistent when emotions weren't.

• Setup isn't the product

  Setup is the sales pitch. The relationship starts afterward. We'd invested heavily in onboarding and underestimated ownership. The redesign reversed that balance.

• Minimal isn't always transparent

  A sparse interface looks elegant in review. When money is moving, users want less decoration and more evidence. Visibility beats minimalism.

• The diagnosis was the design work

  We thought we were solving a conversion problem. We were actually solving a delegation problem. People didn't want the product to invest for them — they wanted it to make consistency easier.