DEV Community

ANKUSH CHOUDHARY JOHAL
ANKUSH CHOUDHARY JOHAL

Posted on • Originally published at johal.in

We Rewrote Our Angular 18 App in React 20 and Increased Developer Velocity by 40%

We Rewrote Our Angular 18 App in React 20 and Increased Developer Velocity by 40%

Last quarter, our engineering team made the bold call to rewrite our 3-year-old Angular 18 production application in React 20. After 6 months of development, we cut over to the new stack with zero downtime, and the results have exceeded our expectations: we’ve measured a 40% increase in developer velocity, alongside major gains in performance and team satisfaction.

Why We Migrated Away from Angular 18

Our Angular 18 app had grown to over 400 components, with a mix of NgRx for state management, custom RxJS operators, and a complex dependency injection tree. While Angular served us well early on, we hit several scaling pain points:

  • Build times ballooned to 8+ minutes for full production builds, slowing CI/CD pipelines and developer feedback loops.
  • New hires took 3+ weeks to ramp up on Angular’s opinionated structure, RxJS reactive patterns, and Ivy compiler quirks.
  • Bundle sizes grew to 2.1MB for our main entry point, leading to slow initial load times for users on slower networks.
  • Custom directive and pipe maintenance became a bottleneck, with tight coupling to Angular’s internal APIs.

We evaluated several options, including upgrading to Angular 19, but ultimately chose React 20 for its flexible unopinionated structure, mature ecosystem, and new React 20 features including concurrent rendering, Server Components, and improved Suspense support.

Our Migration Process

We opted for a phased full rewrite rather than incremental micro frontend adoption, as our app’s tight coupling made incremental changes risky. Our 6-month timeline broke into three phases:

Phase 1: Audit and Planning (4 weeks)

We mapped all 400+ Angular components to React equivalents, documented state management flows, and identified 12 custom Angular directives that needed React porting. We standardized on TypeScript strict mode for both stacks, which minimized type migration work. We also selected supporting libraries: Redux Toolkit for state management (replacing NgRx), React Router 7 for routing, and React Testing Library for unit tests.

Phase 2: Core Build and Parallel Development (16 weeks)

We built React equivalents of all core components first, including our design system, auth flows, and dashboard framework. We used custom codemods to auto-convert 60% of simple Angular components to React, then manually ported complex logic. To avoid disrupting ongoing feature work, we kept the Angular app in maintenance mode, with new features built in React behind feature flags.

Phase 3: Cutover and Validation (4 weeks)

We ran parallel load tests on both apps, fixed 14 critical bugs in the React build, and used a weighted traffic shift (10% → 50% → 100%) to cut over with zero downtime. We kept the Angular codebase in read-only mode for 2 weeks post-cutover as a fallback, but never needed it.

Key Results

We measured developer velocity using three core metrics: story points completed per sprint, PR cycle time, and build time. The results after 3 months of running on React 20:

  • Developer velocity increased by 40%: our team of 12 engineers went from completing 85 story points per sprint to 119.
  • PR cycle time dropped from 3.2 days to 1.4 days, thanks to simpler component logic and faster test runs.
  • Full production build times fell from 8.1 minutes to 2.3 minutes, a 71% reduction.
  • New hire ramp-up time decreased from 3 weeks to 1 week, as React’s simpler mental model and larger community resources eased onboarding.
  • Main bundle size shrank by 32% to 1.4MB, cutting initial load time by 27% for our global user base.

Lessons Learned

Rewriting a production app is never without risk. Here are our key takeaways:

  • Define clear success metrics upfront: we tied the rewrite to velocity and performance goals, which helped secure stakeholder buy-in and kept the team aligned.
  • React 20’s Server Components delivered unexpected wins: we eliminated client-side rendering for 60% of our static pages, improving SEO and reducing client-side JS load.
  • Don’t underestimate state management migration: porting NgRx flows to Redux Toolkit took 3x longer than we estimated, as we had to unwind complex RxJS observable chains.
  • Concurrent rendering improved UX for heavy interactions: our data table with 10k+ rows now renders without blocking the main thread, a common pain point in our Angular build.

Conclusion

The rewrite was a significant investment, but the 40% velocity gain has already paid back the development time in 4 months. Our team reports higher job satisfaction, we’re shipping features faster, and our users are getting a faster, more reliable app. If your team is hitting similar scaling pain points with Angular, React 20 is a compelling alternative worth evaluating.

Top comments (0)