Case study
Flutter app stabilized, redesigned, and fully automated for a federally regulated event trading exchange
The client inherited a Flutter app where every change risked a crash. Wednesday rebuilt the architecture, built a design system, and automated both platform releases without touching the live trading flow.
Fintech / Regulated Exchange · Venture-backed, federally regulated · United States
The challenge
A crash during a live trade is not a UX problem
The client runs a federally regulated exchange where users trade real money on the outcome of events. The product is precise by design. Federal regulation requires it. The app experience had to match that standard.
The team inherited a Flutter app from a previous vendor. The architecture was not suited to the product's complexity. Adding a new market type or a new funding method carried a meaningful risk of breaking the existing order flow. Unexpected errors were common. Crashes reached live users. Every change was a gamble.
The release process was manual and slow. Each improvement waited for a human to build, test, and push to the stores. Iteration speed was constrained by the process, not the engineering.
The app also lacked a design system. Each screen had been styled independently, with inconsistent components that drifted from the brand over time. Maintaining visual consistency required touching every screen individually when anything changed.
A test-driven quality standard was needed. No mechanism existed to catch regressions before they reached users.
“We needed to fix the foundation without breaking what users were already depending on. Every change needed to be predictable.”
The approach
Architecture rebuilt. Design system built. Releases automated.
Wednesday assigned a team of Flutter developers to assess the inherited app and build an implementation plan before touching any live code.
Architecture rebuild. The team updated the app's architecture to isolate features from one another. Adding a new market type no longer risks breaking the funding flow. Updating funding options no longer affects order display. Unexpected errors and crashes dropped dramatically after the update. New functionality could be added with confidence.
Design system. The team built a modular component library from the client's brand guidelines. Every screen now draws from the same set of components. Changing a component propagates consistently throughout the app without per-screen adjustments. The brand is expressed uniformly from the market listing to the trade confirmation screen.
CI pipeline. An automated pipeline runs linting and tests on every change before it can be merged. Nothing reaches the next step without clearing the quality gate. This is the regression floor.
CD pipeline. The build and distribution process is fully automated for both platforms. Android builds go to AppCenter. iOS builds go to TestFlight. Approved changes ship on schedule without manual steps. Release bottlenecks no longer exist.
“Wednesday's team is very methodical in their approach. They never stray from the fundamentals of building software well, and they always think through every step.”
The results
Crashes eliminated. Releases automated. Design consistent.
The architecture rebuild eliminated the unexpected errors and crashes that were routine in the inherited app. Adding new markets, new funding options, and new features no longer carries the risk of breaking the live trading flow.
The design system covers every screen. The brand is expressed consistently throughout the app without per-screen maintenance. When brand tokens change, the change propagates automatically.
Both platforms build and ship automatically. The release process is not a bottleneck. Changes ship on the team's schedule, not on the release manager's availability.
The automated quality gate catches regressions before they reach users. Every change goes through the same checks before it can ship.
“The app is much better now than when we started. They delivered on time, exceeded expectations, and found issues we didn't even know we had.”
ROI
An unreliable trading app costs more than a rebuild. Every crash during a live trade is a support ticket, a potential refund, and a user who does not come back. The architecture rebuild was the cheaper option.
Run the numbers
See what these results would look like for your team size and budget.
“Wednesday's team is very methodical in their approach. They have a unique style of working and never stray from the fundamentals.”
Founder — Fintech platform
Next step
Facing a similar mobile app quality problem?
30 minutes with an engineer. Bring your current setup and your deadline. You leave with a squad shape and a written burn estimate.