Performance glossaryTime to First Byte (TTFB) is the time from requesting a page until the first byte of the response arrives. It reflects server and network speed.
Time to First Byte (TTFB) measures how long a browser waits, after it starts requesting a page, before the first byte of the response arrives. It is the foundation for every other loading metric: until the HTML starts arriving, the browser cannot discover images, styles, or scripts, and nothing can be shown.
What TTFB includes
As Google measures it for a page load, TTFB covers every step before the response starts:
- any redirects;
- the DNS lookup, to find the server's IP address;
- connecting, including the TCP and TLS handshakes;
- sending the request;
- the server preparing the response (see server response time);
- the first byte traveling back.
So a slow TTFB can come from the network and distance as much as from the server. A visitor in Sydney will see a higher TTFB from a server in London than a visitor in Paris does, however fast the server is.
What a good TTFB is
Google's guidance is a TTFB of 0.8 seconds or less at the 75th percentile of real page loads, with over 1.8 seconds counted as poor. TTFB is not a Core Web Vital, and Google treats the threshold as a rough guide; what matters is that TTFB does not hold back the metrics that do count. A slow TTFB makes a good Largest Contentful Paint hard to reach, because the 2.5-second LCP budget starts before the first byte arrives.
How to measure it
- PageSpeed Insights shows field TTFB from real Chrome users, where a page has enough traffic.
- Chrome DevTools: in the Network panel, select the HTML document and open the Timing tab to see each phase.
- The command line:
curl -o /dev/null -s -w "%{time_starttransfer}\n" https://example.com/prints the time to first byte in seconds. - WebPageTest measures it from many locations.
Comparing a cached page, as a logged-out visitor, with an uncached one shows how much caching is doing.
How to improve it on a WordPress site
- Cache pages. Serving stored HTML instead of building it with PHP is the biggest single improvement for most sites. See caching.
- Cache at the edge. A CDN that caches HTML serves it from near the visitor, cutting the network part too.
- Remove redirects from internal links, and make sure visitors land directly on the final HTTPS address.
- Speed up uncached pages with a persistent object cache, a current PHP version, and fewer heavy plugins.
- Host close to the audience, with enough resources for the traffic.
- Use modern protocols. TLS 1.3 and HTTP/3 shorten connection setup.
Source: web.dev: Time to First Byte.