What Is TTFB and How to Reduce Time to First Byte

Time to First Byte is a fundamental web performance diagnostic metric that quantifies the responsiveness of your origin infrastructure and network transit. According to web.dev guidance on optimizing TTFB, TTFB encompasses three distinct sequential phases: (1) Network request propagation and DNS resolution, (2) Server-side processing and backend database execution, and (3) Initial HTTP response header transmission over the wire.

For a concise definition of metric boundaries, see our TTFB glossary entry. In our broader how to improve page speed pillar guide, we identify TTFB as the absolute performance ceiling: if your server takes 700 milliseconds to transmit the first byte, your page cannot paint content faster than 700 milliseconds under any conditions.

According to Chrome Web Vitals standards, TTFB ratings are categorized into three distinct performance tiers evaluated at the 75th percentile of real-world page visits:

While Google specifies 800 milliseconds as the boundary for 'Good', modern websites operating on edge networks routinely achieve TTFB values between 50 and 150 milliseconds. Because Largest Contentful Paint (LCP) includes TTFB as its first sub-part, every 100 milliseconds shaved from server response time translates directly into a 100-millisecond reduction in LCP.

Reducing TTFB requires addressing both network transport latency and origin compute overhead. The following four strategies eliminate the most severe server-side bottlenecks:

When an overseas visitor accesses a central server in North America, each TCP and TLS packet must cross physical oceans, adding 150–300 milliseconds of round-trip transit time (RTT). By routing traffic through an Edge CDN (such as Cloudflare, Fastly, or Vercel Edge Network), the TLS cryptographic handshake is terminated at a local point of presence mere milliseconds from the user.

// Configuring stale-while-revalidate edge caching headers in Next.js
export async function GET() {
  return new Response(JSON.stringify({ status: "ok" }), {
    headers: {
      "Content-Type": "application/json",
      // Serve cached content immediately; revalidate in background
      "Cache-Control": "public, max-age=60, s-maxage=3600, stale-while-revalidate=86400",
      "CDN-Cache-Control": "public, s-maxage=3600",
    },
  });
}

In traditional monolithic SSR, the server must wait for all database queries and external microservices to finish before returning a single byte of HTML. With streaming SSR and React Suspense, the server immediately flushes the document <head>, CSS links, and page skeleton within 30–50 milliseconds. The browser can preload fonts and hero images while secondary dynamic blocks stream in as their data resolves.

Unindexed database queries are a primary culprit behind inflated server latency. If rendering a single dashboard or category route triggers 25 separate database roundtrips (the classic N+1 problem), server response time will degrade exponentially under concurrent traffic. Always audit database query execution plans, create indexes on foreign keys, and memoize query responses using Redis or in-memory LRU caches.

HTTP/3 operates over QUIC, a modern UDP-based transport protocol standardized in IETF RFC 9114. Unlike TCP, which suffers from head-of-line blocking whenever a single packet drops on mobile networks, QUIC multiplexes streams independently and supports zero-round-trip resumption (0-RTT), accelerating reconnection times for returning visitors.

Measuring TTFB requires isolating network transmission from server computation. The following tools provide accurate diagnostic data without synthetic distortion:

• Chrome DevTools Network Tab: The 'Waiting for server response' bar in the Waterfall tab reflects exact TTFB for each individual network request.
• MDN PerformanceResourceTiming API: Programmatically capture responseStart - requestStart in real-user monitoring (RUM) scripts across real human sessions.
• cURL Command Line: The most reliable method to test origin server performance without browser extension interference or client rendering overhead.
• ScanWise Performance Audits: Automatically benchmarks TTFB across multiple pages, highlighting server latency bottlenecks alongside Core Web Vitals.

• web.dev: Optimize Time to First Byte — Google's technical guide to server response times and Web Vitals.
• MDN Web Docs: PerformanceResourceTiming.responseStart — W3C timing specification.
• IETF RFC 9114: HTTP/3 Protocol Specification — Technical standard for QUIC transport and multiplexing.
• Google Search Central: Understanding Core Web Vitals — Search ranking documentation.