Several components are rendering: Client performance at Slack-scale

Performance culture cannot rely on “hero engineering” - it needs to be distributed across the organization through education, tooling, and guardrails

Browser performance constraints give frontend teams only 16ms per frame (60fps) to complete all JavaScript execution, rendering, and styling

Redux’s global state updates caused unnecessary re-renders across components - implementing finer-grained subscriptions helped solve this without completely replacing the architecture

Metrics and observability were crucial for understanding performance issues and getting buy-in - key press lag became their flagship metric

Performance improvements came from understanding systemic issues through investigation rather than just profiling and fixing individual slow functions

Paper cuts (many small performance issues) compound over time and are harder to catch with standard regression testing

Engineers want to write performant code but need the right tools, knowledge and culture to succeed

Long-lived desktop apps like Slack face unique performance challenges compared to typical web pages due to constant updates and interactions

Performance work requires both technical solutions and organizational change through education and evangelism

Building a performance-conscious engineering culture is more sustainable than one-off optimization efforts or complete rewrites